03_M04L01_Lab_Building Autoencoders__M.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p style=\"text-align:center\">\n",
    "    <a href=\"https://skills.network\" target=\"_blank\">\n",
    "    <img src=\"https://cf-courses-data.s3.us.cloud-object-storage.appdomain.cloud/assets/logos/SN_web_lightmode.png\" width=\"200\" alt=\"Skills Network Logo\"  />\n",
    "    </a>\n",
    "</p>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Lab: Building Autoencoders**\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Estimated time needed: **30** minutes\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this lab, you will learn how to build autoencoders using Keras.  \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Learning Objectives\n",
    "\n",
    "By the end of this lab, you will: \n",
    "\n",
    "- Load and preprocess the MNIST dataset for training an autoencoder. \n",
    "\n",
    "- Construct a simple autoencoder model using the Keras functional API. \n",
    "\n",
    "- Train the autoencoder on the MNIST dataset. \n",
    "\n",
    "- Evaluate the performance of the trained autoencoder. \n",
    "\n",
    "- Fine-tune the autoencoder to improve its performance. \n",
    "\n",
    "- Use the autoencoder to denoise images. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step-by-Step Instructions: \n",
    "\n",
    "#### Step 1: Data Preprocessing \n",
    "\n",
    "This exercise prepares the MNIST dataset for training by normalizing the pixel values and flattening the images. Normalization helps in faster convergence during training, and flattening is required because the input layer of our autoencoder expects a one-dimensional vector. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting tensorflow==2.16.2\n",
      "  Downloading tensorflow-2.16.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (4.2 kB)\n",
      "Collecting absl-py>=1.0.0 (from tensorflow==2.16.2)\n",
      "  Downloading absl_py-2.1.0-py3-none-any.whl.metadata (2.3 kB)\n",
      "Collecting astunparse>=1.6.0 (from tensorflow==2.16.2)\n",
      "  Downloading astunparse-1.6.3-py2.py3-none-any.whl.metadata (4.4 kB)\n",
      "Collecting flatbuffers>=23.5.26 (from tensorflow==2.16.2)\n",
      "  Downloading flatbuffers-24.3.25-py2.py3-none-any.whl.metadata (850 bytes)\n",
      "Collecting gast!=0.5.0,!=0.5.1,!=0.5.2,>=0.2.1 (from tensorflow==2.16.2)\n",
      "  Downloading gast-0.6.0-py3-none-any.whl.metadata (1.3 kB)\n",
      "Collecting google-pasta>=0.1.1 (from tensorflow==2.16.2)\n",
      "  Downloading google_pasta-0.2.0-py3-none-any.whl.metadata (814 bytes)\n",
      "Collecting h5py>=3.10.0 (from tensorflow==2.16.2)\n",
      "  Downloading h5py-3.12.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (2.5 kB)\n",
      "Collecting libclang>=13.0.0 (from tensorflow==2.16.2)\n",
      "  Downloading libclang-18.1.1-py2.py3-none-manylinux2010_x86_64.whl.metadata (5.2 kB)\n",
      "Collecting ml-dtypes~=0.3.1 (from tensorflow==2.16.2)\n",
      "  Downloading ml_dtypes-0.3.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (20 kB)\n",
      "Collecting opt-einsum>=2.3.2 (from tensorflow==2.16.2)\n",
      "  Downloading opt_einsum-3.4.0-py3-none-any.whl.metadata (6.3 kB)\n",
      "Requirement already satisfied: packaging in /opt/conda/lib/python3.11/site-packages (from tensorflow==2.16.2) (24.0)\n",
      "Collecting protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 (from tensorflow==2.16.2)\n",
      "  Downloading protobuf-4.25.5-cp37-abi3-manylinux2014_x86_64.whl.metadata (541 bytes)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in /opt/conda/lib/python3.11/site-packages (from tensorflow==2.16.2) (2.31.0)\n",
      "Requirement already satisfied: setuptools in /opt/conda/lib/python3.11/site-packages (from tensorflow==2.16.2) (69.5.1)\n",
      "Requirement already satisfied: six>=1.12.0 in /opt/conda/lib/python3.11/site-packages (from tensorflow==2.16.2) (1.16.0)\n",
      "Collecting termcolor>=1.1.0 (from tensorflow==2.16.2)\n",
      "  Downloading termcolor-2.5.0-py3-none-any.whl.metadata (6.1 kB)\n",
      "Requirement already satisfied: typing-extensions>=3.6.6 in /opt/conda/lib/python3.11/site-packages (from tensorflow==2.16.2) (4.12.2)\n",
      "Collecting wrapt>=1.11.0 (from tensorflow==2.16.2)\n",
      "  Downloading wrapt-1.17.0-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.4 kB)\n",
      "Collecting grpcio<2.0,>=1.24.3 (from tensorflow==2.16.2)\n",
      "  Downloading grpcio-1.68.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (3.9 kB)\n",
      "Collecting tensorboard<2.17,>=2.16 (from tensorflow==2.16.2)\n",
      "  Downloading tensorboard-2.16.2-py3-none-any.whl.metadata (1.6 kB)\n",
      "Collecting keras>=3.0.0 (from tensorflow==2.16.2)\n",
      "  Downloading keras-3.7.0-py3-none-any.whl.metadata (5.8 kB)\n",
      "Collecting tensorflow-io-gcs-filesystem>=0.23.1 (from tensorflow==2.16.2)\n",
      "  Downloading tensorflow_io_gcs_filesystem-0.37.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (14 kB)\n",
      "Collecting numpy<2.0.0,>=1.23.5 (from tensorflow==2.16.2)\n",
      "  Downloading numpy-1.26.4-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (61 kB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m61.0/61.0 kB\u001b[0m \u001b[31m7.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hRequirement already satisfied: wheel<1.0,>=0.23.0 in /opt/conda/lib/python3.11/site-packages (from astunparse>=1.6.0->tensorflow==2.16.2) (0.43.0)\n",
      "Collecting rich (from keras>=3.0.0->tensorflow==2.16.2)\n",
      "  Downloading rich-13.9.4-py3-none-any.whl.metadata (18 kB)\n",
      "Collecting namex (from keras>=3.0.0->tensorflow==2.16.2)\n",
      "  Downloading namex-0.0.8-py3-none-any.whl.metadata (246 bytes)\n",
      "Collecting optree (from keras>=3.0.0->tensorflow==2.16.2)\n",
      "  Downloading optree-0.13.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (47 kB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m47.8/47.8 kB\u001b[0m \u001b[31m7.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.11/site-packages (from requests<3,>=2.21.0->tensorflow==2.16.2) (3.3.2)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.11/site-packages (from requests<3,>=2.21.0->tensorflow==2.16.2) (3.7)\n",
      "Requirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.11/site-packages (from requests<3,>=2.21.0->tensorflow==2.16.2) (2.2.1)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.11/site-packages (from requests<3,>=2.21.0->tensorflow==2.16.2) (2024.8.30)\n",
      "Collecting markdown>=2.6.8 (from tensorboard<2.17,>=2.16->tensorflow==2.16.2)\n",
      "  Downloading Markdown-3.7-py3-none-any.whl.metadata (7.0 kB)\n",
      "Collecting tensorboard-data-server<0.8.0,>=0.7.0 (from tensorboard<2.17,>=2.16->tensorflow==2.16.2)\n",
      "  Downloading tensorboard_data_server-0.7.2-py3-none-manylinux_2_31_x86_64.whl.metadata (1.1 kB)\n",
      "Collecting werkzeug>=1.0.1 (from tensorboard<2.17,>=2.16->tensorflow==2.16.2)\n",
      "  Downloading werkzeug-3.1.3-py3-none-any.whl.metadata (3.7 kB)\n",
      "Requirement already satisfied: MarkupSafe>=2.1.1 in /opt/conda/lib/python3.11/site-packages (from werkzeug>=1.0.1->tensorboard<2.17,>=2.16->tensorflow==2.16.2) (2.1.5)\n",
      "Collecting markdown-it-py>=2.2.0 (from rich->keras>=3.0.0->tensorflow==2.16.2)\n",
      "  Downloading markdown_it_py-3.0.0-py3-none-any.whl.metadata (6.9 kB)\n",
      "Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /opt/conda/lib/python3.11/site-packages (from rich->keras>=3.0.0->tensorflow==2.16.2) (2.18.0)\n",
      "Collecting mdurl~=0.1 (from markdown-it-py>=2.2.0->rich->keras>=3.0.0->tensorflow==2.16.2)\n",
      "  Downloading mdurl-0.1.2-py3-none-any.whl.metadata (1.6 kB)\n",
      "Downloading tensorflow-2.16.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (590.7 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m590.7/590.7 MB\u001b[0m \u001b[31m871.9 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading absl_py-2.1.0-py3-none-any.whl (133 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m133.7/133.7 kB\u001b[0m \u001b[31m19.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading astunparse-1.6.3-py2.py3-none-any.whl (12 kB)\n",
      "Downloading flatbuffers-24.3.25-py2.py3-none-any.whl (26 kB)\n",
      "Downloading gast-0.6.0-py3-none-any.whl (21 kB)\n",
      "Downloading google_pasta-0.2.0-py3-none-any.whl (57 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m57.5/57.5 kB\u001b[0m \u001b[31m8.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading grpcio-1.68.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.9 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.9/5.9 MB\u001b[0m \u001b[31m118.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading h5py-3.12.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.5 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.5/5.5 MB\u001b[0m \u001b[31m126.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading keras-3.7.0-py3-none-any.whl (1.2 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.2/1.2 MB\u001b[0m \u001b[31m92.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading libclang-18.1.1-py2.py3-none-manylinux2010_x86_64.whl (24.5 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m24.5/24.5 MB\u001b[0m \u001b[31m24.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading ml_dtypes-0.3.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.2 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.2/2.2 MB\u001b[0m \u001b[31m65.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading numpy-1.26.4-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (18.3 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m18.3/18.3 MB\u001b[0m \u001b[31m32.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading opt_einsum-3.4.0-py3-none-any.whl (71 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m71.9/71.9 kB\u001b[0m \u001b[31m9.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading protobuf-4.25.5-cp37-abi3-manylinux2014_x86_64.whl (294 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m294.6/294.6 kB\u001b[0m \u001b[31m34.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading tensorboard-2.16.2-py3-none-any.whl (5.5 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.5/5.5 MB\u001b[0m \u001b[31m66.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m\n",
      "\u001b[?25hDownloading tensorflow_io_gcs_filesystem-0.37.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.1 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.1/5.1 MB\u001b[0m \u001b[31m54.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m:00:01\u001b[0m\n",
      "\u001b[?25hDownloading termcolor-2.5.0-py3-none-any.whl (7.8 kB)\n",
      "Downloading wrapt-1.17.0-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (83 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m83.2/83.2 kB\u001b[0m \u001b[31m12.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading Markdown-3.7-py3-none-any.whl (106 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m106.3/106.3 kB\u001b[0m \u001b[31m12.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading tensorboard_data_server-0.7.2-py3-none-manylinux_2_31_x86_64.whl (6.6 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.6/6.6 MB\u001b[0m \u001b[31m64.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m\n",
      "\u001b[?25hDownloading werkzeug-3.1.3-py3-none-any.whl (224 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m224.5/224.5 kB\u001b[0m \u001b[31m23.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading namex-0.0.8-py3-none-any.whl (5.8 kB)\n",
      "Downloading optree-0.13.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (391 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m391.8/391.8 kB\u001b[0m \u001b[31m43.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading rich-13.9.4-py3-none-any.whl (242 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m242.4/242.4 kB\u001b[0m \u001b[31m34.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading markdown_it_py-3.0.0-py3-none-any.whl (87 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.5/87.5 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading mdurl-0.1.2-py3-none-any.whl (10.0 kB)\n",
      "Installing collected packages: namex, libclang, flatbuffers, wrapt, werkzeug, termcolor, tensorflow-io-gcs-filesystem, tensorboard-data-server, protobuf, optree, opt-einsum, numpy, mdurl, markdown, grpcio, google-pasta, gast, astunparse, absl-py, tensorboard, ml-dtypes, markdown-it-py, h5py, rich, keras, tensorflow\n",
      "Successfully installed absl-py-2.1.0 astunparse-1.6.3 flatbuffers-24.3.25 gast-0.6.0 google-pasta-0.2.0 grpcio-1.68.0 h5py-3.12.1 keras-3.7.0 libclang-18.1.1 markdown-3.7 markdown-it-py-3.0.0 mdurl-0.1.2 ml-dtypes-0.3.2 namex-0.0.8 numpy-1.26.4 opt-einsum-3.4.0 optree-0.13.1 protobuf-4.25.5 rich-13.9.4 tensorboard-2.16.2 tensorboard-data-server-0.7.2 tensorflow-2.16.2 tensorflow-io-gcs-filesystem-0.37.1 termcolor-2.5.0 werkzeug-3.1.3 wrapt-1.17.0\n"
     ]
    }
   ],
   "source": [
    "!pip install tensorflow==2.16.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-11-28 16:14:54.246042: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2024-11-28 16:14:54.248415: I external/local_tsl/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
      "2024-11-28 16:14:54.257409: I external/local_tsl/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
      "2024-11-28 16:14:54.270034: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:479] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
      "2024-11-28 16:14:54.294011: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:10575] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
      "2024-11-28 16:14:54.294071: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1442] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
      "2024-11-28 16:14:54.309693: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2024-11-28 16:14:55.439002: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
      "\u001b[1m11490434/11490434\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 0us/step\n"
     ]
    }
   ],
   "source": [
    "import numpy as np \n",
    "from tensorflow.keras.datasets import mnist \n",
    "\n",
    "# Load the dataset \n",
    "(x_train, _), (x_test, _) = mnist.load_data() \n",
    "\n",
    "# Normalize the pixel values \n",
    "x_train = x_train.astype('float32') / 255. \n",
    "x_test = x_test.astype('float32') / 255. \n",
    "\n",
    "# Flatten the images \n",
    "x_train = x_train.reshape((len(x_train), np.prod(x_train.shape[1:]))) \n",
    "x_test = x_test.reshape((len(x_test), np.prod(x_test.shape[1:]))) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "- Use Keras to load the MNIST dataset. \n",
    "- Normalize the image pixel values to the range [0, 1]. \n",
    "- Flatten the 28x28 images to a 784-dimensional vector to reshape the data. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Step 2: Building the Autoencoder Model \n",
    "\n",
    "This exercise involves building an autoencoder with an encoder that compresses the input to 32 dimensions and a decoder that reconstructs the input from these 32 dimensions. The model is compiled with the Adam optimizer and binary crossentropy loss. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ input_layer (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">784</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)             │        <span style=\"color: #00af00; text-decoration-color: #00af00\">50,240</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)             │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,080</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)             │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,112</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">784</span>)            │        <span style=\"color: #00af00; text-decoration-color: #00af00\">50,960</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ input_layer (\u001b[38;5;33mInputLayer\u001b[0m)        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense (\u001b[38;5;33mDense\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m)             │        \u001b[38;5;34m50,240\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m)             │         \u001b[38;5;34m2,080\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_2 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m)             │         \u001b[38;5;34m2,112\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │        \u001b[38;5;34m50,960\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">105,392</span> (411.69 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m105,392\u001b[0m (411.69 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">105,392</span> (411.69 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m105,392\u001b[0m (411.69 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from tensorflow.keras.models import Model \n",
    "from tensorflow.keras.layers import Input, Dense \n",
    "\n",
    "# Encoder \n",
    "input_layer = Input(shape=(784,)) \n",
    "encoded = Dense(64, activation='relu')(input_layer) \n",
    "\n",
    "# Bottleneck \n",
    "bottleneck = Dense(32, activation='relu')(encoded) \n",
    "\n",
    "# Decoder \n",
    "decoded = Dense(64, activation='relu')(bottleneck) \n",
    "output_layer = Dense(784, activation='sigmoid')(decoded) \n",
    "\n",
    "# Autoencoder model \n",
    "autoencoder = Model(input_layer, output_layer) \n",
    "\n",
    "# Compile the model \n",
    "autoencoder.compile(optimizer='adam', loss='binary_crossentropy') \n",
    "\n",
    "# Summary of the model \n",
    "autoencoder.summary() "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "\n",
    "**1. Define the Encoder:**\n",
    "- Create an input layer with 784 neurons. \n",
    "- Add a Dense layer with 64 neurons and ReLU activation. \n",
    "\n",
    "**2. Define the Bottleneck:**\n",
    "- Add a Dense layer with 32 neurons and ReLU activation. \n",
    "\n",
    "**3. Define the Decoder:**\n",
    "- Add a Dense layer with 64 neurons and ReLU activation. \n",
    "- Add an output layer with 784 neurons and sigmoid activation. \n",
    "\n",
    "**4. Compile the Model:**\n",
    "- Use the Adam optimizer and binary crossentropy loss.  \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Step 3: Training the Autoencoder \n",
    "\n",
    "In this exercise, the autoencoder is trained to reconstruct the MNIST images. The training data is both the input and the target, as the autoencoder learns to map the input to itself. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 30ms/step - loss: 0.3678 - val_loss: 0.1746\n",
      "Epoch 2/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 31ms/step - loss: 0.1654 - val_loss: 0.1427\n",
      "Epoch 3/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1397 - val_loss: 0.1273\n",
      "Epoch 4/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1268 - val_loss: 0.1197\n",
      "Epoch 5/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1200 - val_loss: 0.1144\n",
      "Epoch 6/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 26ms/step - loss: 0.1152 - val_loss: 0.1113\n",
      "Epoch 7/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1121 - val_loss: 0.1084\n",
      "Epoch 8/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1093 - val_loss: 0.1068\n",
      "Epoch 9/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1073 - val_loss: 0.1042\n",
      "Epoch 10/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1050 - val_loss: 0.1023\n",
      "Epoch 11/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1035 - val_loss: 0.1009\n",
      "Epoch 12/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1020 - val_loss: 0.1000\n",
      "Epoch 13/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1010 - val_loss: 0.0991\n",
      "Epoch 14/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1001 - val_loss: 0.0983\n",
      "Epoch 15/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0992 - val_loss: 0.0977\n",
      "Epoch 16/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0987 - val_loss: 0.0970\n",
      "Epoch 17/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.0979 - val_loss: 0.0963\n",
      "Epoch 18/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0973 - val_loss: 0.0956\n",
      "Epoch 19/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0965 - val_loss: 0.0948\n",
      "Epoch 20/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.0961 - val_loss: 0.0943\n",
      "Epoch 21/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0954 - val_loss: 0.0938\n",
      "Epoch 22/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0949 - val_loss: 0.0935\n",
      "Epoch 23/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0945 - val_loss: 0.0932\n",
      "Epoch 24/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0945 - val_loss: 0.0931\n",
      "Epoch 25/25\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.0941 - val_loss: 0.0929\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x7f327d5b8c10>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "autoencoder.fit(\n",
    "    x_train, x_train,  \n",
    "    epochs=25,  \n",
    "    batch_size=256,  \n",
    "    shuffle=True,  \n",
    "    validation_data=(x_test, x_test)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "- Use the `fit` method to train the model on the training data. \n",
    "- Set the number of epochs to 25 and the batch size to 256.. \n",
    "- Use the test data for validation. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Step 4: Evaluating the Autoencoder \n",
    "\n",
    "This exercise evaluates the autoencoder by reconstructing the test images and comparing them to the original images. Visualization helps in understanding how well the autoencoder has learned to reconstruct the input data. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting matplotlib==3.9.2\n",
      "  Downloading matplotlib-3.9.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (11 kB)\n",
      "Collecting contourpy>=1.0.1 (from matplotlib==3.9.2)\n",
      "  Downloading contourpy-1.3.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (5.4 kB)\n",
      "Collecting cycler>=0.10 (from matplotlib==3.9.2)\n",
      "  Downloading cycler-0.12.1-py3-none-any.whl.metadata (3.8 kB)\n",
      "Collecting fonttools>=4.22.0 (from matplotlib==3.9.2)\n",
      "  Downloading fonttools-4.55.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (164 kB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m164.5/164.5 kB\u001b[0m \u001b[31m19.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hCollecting kiwisolver>=1.3.1 (from matplotlib==3.9.2)\n",
      "  Downloading kiwisolver-1.4.7-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.3 kB)\n",
      "Requirement already satisfied: numpy>=1.23 in /opt/conda/lib/python3.11/site-packages (from matplotlib==3.9.2) (1.26.4)\n",
      "Requirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.11/site-packages (from matplotlib==3.9.2) (24.0)\n",
      "Collecting pillow>=8 (from matplotlib==3.9.2)\n",
      "  Downloading pillow-11.0.0-cp311-cp311-manylinux_2_28_x86_64.whl.metadata (9.1 kB)\n",
      "Collecting pyparsing>=2.3.1 (from matplotlib==3.9.2)\n",
      "  Downloading pyparsing-3.2.0-py3-none-any.whl.metadata (5.0 kB)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /opt/conda/lib/python3.11/site-packages (from matplotlib==3.9.2) (2.9.0)\n",
      "Requirement already satisfied: six>=1.5 in /opt/conda/lib/python3.11/site-packages (from python-dateutil>=2.7->matplotlib==3.9.2) (1.16.0)\n",
      "Downloading matplotlib-3.9.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (8.3 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m8.3/8.3 MB\u001b[0m \u001b[31m101.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading contourpy-1.3.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (326 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m326.2/326.2 kB\u001b[0m \u001b[31m37.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading cycler-0.12.1-py3-none-any.whl (8.3 kB)\n",
      "Downloading fonttools-4.55.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (4.9 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.9/4.9 MB\u001b[0m \u001b[31m96.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m\n",
      "\u001b[?25hDownloading kiwisolver-1.4.7-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.4 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.4/1.4 MB\u001b[0m \u001b[31m74.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hDownloading pillow-11.0.0-cp311-cp311-manylinux_2_28_x86_64.whl (4.4 MB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.4/4.4 MB\u001b[0m \u001b[31m101.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m\n",
      "\u001b[?25hDownloading pyparsing-3.2.0-py3-none-any.whl (106 kB)\n",
      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m106.9/106.9 kB\u001b[0m \u001b[31m13.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hInstalling collected packages: pyparsing, pillow, kiwisolver, fonttools, cycler, contourpy, matplotlib\n",
      "Successfully installed contourpy-1.3.1 cycler-0.12.1 fonttools-4.55.0 kiwisolver-1.4.7 matplotlib-3.9.2 pillow-11.0.0 pyparsing-3.2.0\n"
     ]
    }
   ],
   "source": [
    "!pip install matplotlib==3.9.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt \n",
    "\n",
    "# Predict the test data \n",
    "reconstructed = autoencoder.predict(x_test) \n",
    "\n",
    "# Visualize the results \n",
    "n = 10  # Number of digits to display \n",
    "plt.figure(figsize=(20, 4)) \n",
    "\n",
    "for i in range(n): \n",
    "    # Display original \n",
    "    ax = plt.subplot(2, n, i + 1) \n",
    "    plt.imshow(x_test[i].reshape(28, 28)) \n",
    "    plt.gray() \n",
    "    ax.get_xaxis().set_visible(False) \n",
    "    ax.get_yaxis().set_visible(False) \n",
    "\n",
    "    # Display reconstruction \n",
    "    ax = plt.subplot(2, n, i + 1 + n) \n",
    "    plt.imshow(reconstructed[i].reshape(28, 28)) \n",
    "    plt.gray() \n",
    "    ax.get_xaxis().set_visible(False) \n",
    "    ax.get_yaxis().set_visible(False) \n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "\n",
    "**1. Reconstruct Images:**\n",
    "- Use the autoencoder to predict the test data. \n",
    "- Compare the original test images with the reconstructed images. \n",
    "\n",
    "**2. Visualize the Results:**\n",
    "- Plot a few examples of original and reconstructed images side by side. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Step 5: Fine-Tuning the Autoencoder \n",
    "\n",
    "Fine-tuning the autoencoder by unfreezing some layers can help in improving its performance. In this exercise, you unfreeze the last four layers and train the model again for a few more epochs.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 28ms/step - loss: 0.0948 - val_loss: 0.0925\n",
      "Epoch 2/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0937 - val_loss: 0.0926\n",
      "Epoch 3/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 29ms/step - loss: 0.0936 - val_loss: 0.0925\n",
      "Epoch 4/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0935 - val_loss: 0.0922\n",
      "Epoch 5/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0934 - val_loss: 0.0923\n",
      "Epoch 6/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0930 - val_loss: 0.0921\n",
      "Epoch 7/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0931 - val_loss: 0.0920\n",
      "Epoch 8/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0929 - val_loss: 0.0920\n",
      "Epoch 9/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0931 - val_loss: 0.0919\n",
      "Epoch 10/10\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0926 - val_loss: 0.0917\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.history.History at 0x7f327d64b8d0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Unfreeze the top layers of the encoder\n",
    "for layer in autoencoder.layers[-4:]: \n",
    "    layer.trainable = True \n",
    "\n",
    "# Compile the model again\n",
    "autoencoder.compile(optimizer='adam', loss='binary_crossentropy') \n",
    "\n",
    "# Train the model again\n",
    "autoencoder.fit(x_train, x_train,  \n",
    "                epochs=10,  \n",
    "                batch_size=256,  \n",
    "                shuffle=True,  \n",
    "                validation_data=(x_test, x_test))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "\n",
    "**1. Unfreeze the Encoder Layers:**\n",
    "- Unfreeze the last four layers of the encoder. \n",
    "\n",
    "**2. Compile and Train the Model:**\n",
    "- Recompile the model. \n",
    "- Train the model again for 10 epochs with the same training and validation data.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Step 6: Denoising Images with Autoencoder \n",
    "\n",
    "In this exercise, you add random noise to the dataset and train the autoencoder to denoise the images. The autoencoder learns to reconstruct the original images from the noisy input, which can be visualized by comparing the noisy, denoised, and original images. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 44ms/step - loss: 0.2103 - val_loss: 0.1434\n",
      "Epoch 2/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1415 - val_loss: 0.1340\n",
      "Epoch 3/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1335 - val_loss: 0.1296\n",
      "Epoch 4/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1298 - val_loss: 0.1273\n",
      "Epoch 5/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1277 - val_loss: 0.1258\n",
      "Epoch 6/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1262 - val_loss: 0.1250\n",
      "Epoch 7/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1254 - val_loss: 0.1246\n",
      "Epoch 8/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 41ms/step - loss: 0.1247 - val_loss: 0.1239\n",
      "Epoch 9/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 45ms/step - loss: 0.1244 - val_loss: 0.1240\n",
      "Epoch 10/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1238 - val_loss: 0.1245\n",
      "Epoch 11/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1240 - val_loss: 0.1229\n",
      "Epoch 12/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1236 - val_loss: 0.1229\n",
      "Epoch 13/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1234 - val_loss: 0.1236\n",
      "Epoch 14/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 41ms/step - loss: 0.1230 - val_loss: 0.1224\n",
      "Epoch 15/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1229 - val_loss: 0.1226\n",
      "Epoch 16/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 43ms/step - loss: 0.1228 - val_loss: 0.1222\n",
      "Epoch 17/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1227 - val_loss: 0.1222\n",
      "Epoch 18/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1224 - val_loss: 0.1220\n",
      "Epoch 19/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1226 - val_loss: 0.1220\n",
      "Epoch 20/20\n",
      "\u001b[1m118/118\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 42ms/step - loss: 0.1225 - val_loss: 0.1233\n",
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 30 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Add noise to the data\n",
    "noise_factor = 0.5\n",
    "x_train_noisy = x_train + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_train.shape)\n",
    "x_test_noisy = x_test + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_test.shape)\n",
    "x_train_noisy = np.clip(x_train_noisy, 0., 1.)\n",
    "x_test_noisy = np.clip(x_test_noisy, 0., 1.)\n",
    "\n",
    "# Train the autoencoder with noisy data\n",
    "autoencoder.fit(\n",
    "    x_train_noisy, x_train,\n",
    "    epochs=20,\n",
    "    batch_size=512,\n",
    "    shuffle=True,\n",
    "    validation_data=(x_test_noisy, x_test)\n",
    ")\n",
    "\n",
    "# Denoise the test images\n",
    "reconstructed_noisy = autoencoder.predict(x_test_noisy)\n",
    "\n",
    "# Visualize the results\n",
    "n = 10  # Number of digits to display\n",
    "plt.figure(figsize=(20, 6))\n",
    "for i in range(n):\n",
    "    # Display noisy images\n",
    "    ax = plt.subplot(3, n, i + 1)\n",
    "    plt.imshow(x_test_noisy[i].reshape(28, 28))\n",
    "    plt.gray()\n",
    "    ax.get_xaxis().set_visible(False)\n",
    "    ax.get_yaxis().set_visible(False)\n",
    "    \n",
    "    # Display denoised images\n",
    "    ax = plt.subplot(3, n, i + 1 + n)\n",
    "    plt.imshow(reconstructed_noisy[i].reshape(28, 28))\n",
    "    plt.gray()\n",
    "    ax.get_xaxis().set_visible(False)\n",
    "    ax.get_yaxis().set_visible(False)\n",
    "\n",
    "    # Display original images\n",
    "    ax = plt.subplot(3, n, i + 1 + 2 * n)\n",
    "    plt.imshow(x_test[i].reshape(28, 28))\n",
    "    plt.gray()\n",
    "    ax.get_xaxis().set_visible(False)\n",
    "    ax.get_yaxis().set_visible(False)\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the above code: \n",
    "\n",
    "**1. Add noise to the data:**\n",
    "- Add random noise to the training and test data. \n",
    "- Train the Autoencoder with noisy data: \n",
    "-Train the autoencoder using the noisy images as input and the original images as target. \n",
    "\n",
    "**2. Evaluate the denoising performance:**\n",
    "- Use the autoencoder to denoise the test images. \n",
    "- Compare the noisy, denoised, and original images. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Practice Exercises: \n",
    "\n",
    "### Exercise 1: Exploring Different Bottleneck Sizes \n",
    "\n",
    "#### Objective: \n",
    "\n",
    "To understand the impact of different bottleneck sizes on the performance of the autoencoder. \n",
    "\n",
    "#### Instructions: \n",
    "\n",
    "**1. Define new models with different bottleneck sizes:**\n",
    "- Create three new autoencoder models, each with a different bottleneck size (e.g., 16, 32, and 64 neurons). \n",
    "- Use the same encoder and decoder architecture as in the main lab but change the number of neurons in the bottleneck layer. \n",
    "\n",
    "**2. Train the models:**\n",
    "- Train each model on the MNIST dataset for 50 epochs with a batch size of 256. \n",
    "- Use the same preprocessing steps as in the main lab. \n",
    "\n",
    "**3. Evaluate and Compare the Models:**\n",
    "- Evaluate the performance of each model on the test data. \n",
    "- Compare the reconstruction loss of the models to understand how the bottleneck size affects the autoencoder's ability to reconstruct the input data. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 27ms/step - loss: 0.3738 - val_loss: 0.1926\n",
      "Epoch 2/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.1828 - val_loss: 0.1596\n",
      "Epoch 3/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1574 - val_loss: 0.1462\n",
      "Epoch 4/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1450 - val_loss: 0.1374\n",
      "Epoch 5/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1373 - val_loss: 0.1308\n",
      "Epoch 6/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1310 - val_loss: 0.1269\n",
      "Epoch 7/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1274 - val_loss: 0.1235\n",
      "Epoch 8/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1243 - val_loss: 0.1207\n",
      "Epoch 9/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1213 - val_loss: 0.1186\n",
      "Epoch 10/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1199 - val_loss: 0.1172\n",
      "Epoch 11/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1184 - val_loss: 0.1163\n",
      "Epoch 12/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1175 - val_loss: 0.1156\n",
      "Epoch 13/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1168 - val_loss: 0.1149\n",
      "Epoch 14/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1160 - val_loss: 0.1141\n",
      "Epoch 15/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1152 - val_loss: 0.1134\n",
      "Epoch 16/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1150 - val_loss: 0.1130\n",
      "Epoch 17/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1140 - val_loss: 0.1126\n",
      "Epoch 18/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1139 - val_loss: 0.1121\n",
      "Epoch 19/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1132 - val_loss: 0.1117\n",
      "Epoch 20/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1129 - val_loss: 0.1113\n",
      "Epoch 1/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 27ms/step - loss: 0.3557 - val_loss: 0.1712\n",
      "Epoch 2/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1629 - val_loss: 0.1394\n",
      "Epoch 3/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1372 - val_loss: 0.1255\n",
      "Epoch 4/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1251 - val_loss: 0.1190\n",
      "Epoch 5/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1196 - val_loss: 0.1147\n",
      "Epoch 6/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1155 - val_loss: 0.1106\n",
      "Epoch 7/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1114 - val_loss: 0.1080\n",
      "Epoch 8/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1088 - val_loss: 0.1060\n",
      "Epoch 9/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.1070 - val_loss: 0.1045\n",
      "Epoch 10/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.1055 - val_loss: 0.1028\n",
      "Epoch 11/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1036 - val_loss: 0.1011\n",
      "Epoch 12/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1021 - val_loss: 0.1000\n",
      "Epoch 13/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1011 - val_loss: 0.0989\n",
      "Epoch 14/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1000 - val_loss: 0.0978\n",
      "Epoch 15/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.0990 - val_loss: 0.0971\n",
      "Epoch 16/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 28ms/step - loss: 0.0983 - val_loss: 0.0963\n",
      "Epoch 17/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0972 - val_loss: 0.0958\n",
      "Epoch 18/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0969 - val_loss: 0.0951\n",
      "Epoch 19/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0963 - val_loss: 0.0946\n",
      "Epoch 20/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.0956 - val_loss: 0.0940\n",
      "Epoch 1/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 28ms/step - loss: 0.3592 - val_loss: 0.1692\n",
      "Epoch 2/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1602 - val_loss: 0.1349\n",
      "Epoch 3/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1322 - val_loss: 0.1205\n",
      "Epoch 4/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.1197 - val_loss: 0.1123\n",
      "Epoch 5/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1123 - val_loss: 0.1068\n",
      "Epoch 6/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1074 - val_loss: 0.1036\n",
      "Epoch 7/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.1040 - val_loss: 0.1010\n",
      "Epoch 8/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.1017 - val_loss: 0.0987\n",
      "Epoch 9/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0997 - val_loss: 0.0971\n",
      "Epoch 10/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.0980 - val_loss: 0.0953\n",
      "Epoch 11/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0961 - val_loss: 0.0942\n",
      "Epoch 12/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0953 - val_loss: 0.0931\n",
      "Epoch 13/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0941 - val_loss: 0.0922\n",
      "Epoch 14/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0931 - val_loss: 0.0914\n",
      "Epoch 15/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 25ms/step - loss: 0.0922 - val_loss: 0.0906\n",
      "Epoch 16/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0915 - val_loss: 0.0899\n",
      "Epoch 17/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0909 - val_loss: 0.0895\n",
      "Epoch 18/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0900 - val_loss: 0.0886\n",
      "Epoch 19/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0896 - val_loss: 0.0882\n",
      "Epoch 20/20\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.0891 - val_loss: 0.0877\n",
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - loss: 0.1119\n",
      "Bottleneck size 16 - Test loss: 0.11126761138439178\n",
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - loss: 0.0946\n",
      "Bottleneck size 32 - Test loss: 0.09402798116207123\n",
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - loss: 0.0880\n",
      "Bottleneck size 64 - Test loss: 0.08769935369491577\n"
     ]
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "# Define and train three different autoencoders with varying bottleneck sizes\n",
    "bottleneck_sizes = [16, 32, 64]\n",
    "autoencoders = []\n",
    "\n",
    "for size in bottleneck_sizes:\n",
    "    # Encoder\n",
    "    input_layer = Input(shape=(784,))\n",
    "    encoded = Dense(64, activation='relu')(input_layer)\n",
    "    bottleneck = Dense(size, activation='relu')(encoded)\n",
    "\n",
    "    # Decoder\n",
    "    decoded = Dense(64, activation='relu')(bottleneck)\n",
    "    output_layer = Dense(784, activation='sigmoid')(decoded)\n",
    "\n",
    "    # Autoencoder model\n",
    "    autoencoder = Model(input_layer, output_layer)\n",
    "    autoencoder.compile(optimizer='adam', loss='binary_crossentropy')\n",
    "    autoencoder.fit(\n",
    "        x_train,\n",
    "        x_train,\n",
    "        epochs=20,\n",
    "        batch_size=256,\n",
    "        shuffle=True,\n",
    "        validation_data=(x_test, x_test)\n",
    "    )\n",
    "    autoencoders.append(autoencoder)\n",
    "\n",
    "# Evaluate and compare the models\n",
    "for i, size in enumerate(bottleneck_sizes):\n",
    "    loss = autoencoders[i].evaluate(x_test, x_test)\n",
    "    print(f'Bottleneck size {size} - Test loss: {loss}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "# Define and train three different autoencoders with varying bottleneck sizes\n",
    "bottleneck_sizes = [16, 32, 64]\n",
    "autoencoders = []\n",
    "\n",
    "for size in bottleneck_sizes:\n",
    "    # Encoder\n",
    "    input_layer = Input(shape=(784,))\n",
    "    encoded = Dense(64, activation='relu')(input_layer)\n",
    "    bottleneck = Dense(size, activation='relu')(encoded)\n",
    "\n",
    "    # Decoder\n",
    "    decoded = Dense(64, activation='relu')(bottleneck)\n",
    "    output_layer = Dense(784, activation='sigmoid')(decoded)\n",
    "\n",
    "    # Autoencoder model\n",
    "    autoencoder = Model(input_layer, output_layer)\n",
    "    autoencoder.compile(optimizer='adam', loss='binary_crossentropy')\n",
    "    autoencoder.fit(\n",
    "        x_train,\n",
    "        x_train,\n",
    "        epochs=20,\n",
    "        batch_size=256,\n",
    "        shuffle=True,\n",
    "        validation_data=(x_test, x_test)\n",
    "    )\n",
    "    autoencoders.append(autoencoder)\n",
    "\n",
    "# Evaluate and compare the models\n",
    "for i, size in enumerate(bottleneck_sizes):\n",
    "    loss = autoencoders[i].evaluate(x_test, x_test)\n",
    "    print(f'Bottleneck size {size} - Test loss: {loss}')\n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 2 - Adding Regularization to the Autoencoder \n",
    " \n",
    "#### Objective: \n",
    "\n",
    "To explore the effect of regularization on the performance of the autoencoder. \n",
    "\n",
    "#### Instructions: \n",
    "\n",
    "**1. Modify the model:**\n",
    "- Add L2 regularization to the Dense layers in both the encoder and decoder parts of the autoencoder. \n",
    "\n",
    "**2. Train the model:**\n",
    "- Train the modified autoencoder on the MNIST dataset for 50 epochs with a batch size of 256. \n",
    "\n",
    "**3. Evaluate and compare:**\n",
    "- Evaluate the performance of the regularized autoencoder and compare it with the non-regularized version. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 28ms/step - loss: 1.7676 - val_loss: 0.4408\n",
      "Epoch 2/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.4228 - val_loss: 0.3945\n",
      "Epoch 3/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.3910 - val_loss: 0.3797\n",
      "Epoch 4/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3763 - val_loss: 0.3662\n",
      "Epoch 5/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3636 - val_loss: 0.3545\n",
      "Epoch 6/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.3525 - val_loss: 0.3447\n",
      "Epoch 7/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.3425 - val_loss: 0.3367\n",
      "Epoch 8/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.3353 - val_loss: 0.3300\n",
      "Epoch 9/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.3287 - val_loss: 0.3239\n",
      "Epoch 10/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.3227 - val_loss: 0.3184\n",
      "Epoch 11/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3171 - val_loss: 0.3131\n",
      "Epoch 12/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3121 - val_loss: 0.3083\n",
      "Epoch 13/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3074 - val_loss: 0.3035\n",
      "Epoch 14/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.3029 - val_loss: 0.2991\n",
      "Epoch 15/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.2983 - val_loss: 0.2950\n",
      "Epoch 16/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2940 - val_loss: 0.2912\n",
      "Epoch 17/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.2903 - val_loss: 0.2876\n",
      "Epoch 18/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2868 - val_loss: 0.2842\n",
      "Epoch 19/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2834 - val_loss: 0.2815\n",
      "Epoch 20/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.2813 - val_loss: 0.2787\n",
      "Epoch 21/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2782 - val_loss: 0.2765\n",
      "Epoch 22/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2762 - val_loss: 0.2749\n",
      "Epoch 23/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.2743 - val_loss: 0.2731\n",
      "Epoch 24/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2727 - val_loss: 0.2718\n",
      "Epoch 25/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2719 - val_loss: 0.2708\n",
      "Epoch 26/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2704 - val_loss: 0.2696\n",
      "Epoch 27/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2695 - val_loss: 0.2687\n",
      "Epoch 28/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2689 - val_loss: 0.2679\n",
      "Epoch 29/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2683 - val_loss: 0.2673\n",
      "Epoch 30/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2673 - val_loss: 0.2670\n",
      "Epoch 31/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2675 - val_loss: 0.2663\n",
      "Epoch 32/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2665 - val_loss: 0.2660\n",
      "Epoch 33/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 27ms/step - loss: 0.2660 - val_loss: 0.2656\n",
      "Epoch 34/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 27ms/step - loss: 0.2660 - val_loss: 0.2651\n",
      "Epoch 35/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 26ms/step - loss: 0.2653 - val_loss: 0.2649\n",
      "Epoch 36/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2657 - val_loss: 0.2647\n",
      "Epoch 37/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2646 - val_loss: 0.2647\n",
      "Epoch 38/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2649 - val_loss: 0.2643\n",
      "Epoch 39/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2649 - val_loss: 0.2639\n",
      "Epoch 40/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2643 - val_loss: 0.2639\n",
      "Epoch 41/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2641 - val_loss: 0.2640\n",
      "Epoch 42/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2641 - val_loss: 0.2637\n",
      "Epoch 43/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2638 - val_loss: 0.2637\n",
      "Epoch 44/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 29ms/step - loss: 0.2636 - val_loss: 0.2633\n",
      "Epoch 45/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2638 - val_loss: 0.2634\n",
      "Epoch 46/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2634 - val_loss: 0.2633\n",
      "Epoch 47/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2637 - val_loss: 0.2632\n",
      "Epoch 48/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2636 - val_loss: 0.2631\n",
      "Epoch 49/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 27ms/step - loss: 0.2634 - val_loss: 0.2631\n",
      "Epoch 50/50\n",
      "\u001b[1m235/235\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 28ms/step - loss: 0.2631 - val_loss: 0.2630\n",
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 4ms/step - loss: 0.2576\n",
      "Regularized Autoencoder - Test loss: 0.26302117109298706\n"
     ]
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "from tensorflow.keras.regularizers import l2 \n",
    "\n",
    "# Encoder with L2 regularization \n",
    "input_layer = Input(shape=(784,)) \n",
    "encoded = Dense(64, activation='relu', kernel_regularizer=l2(0.01))(input_layer) \n",
    "bottleneck = Dense(32, activation='relu', kernel_regularizer=l2(0.01))(encoded) \n",
    "\n",
    "# Decoder with L2 regularization \n",
    "decoded = Dense(64, activation='relu', kernel_regularizer=l2(0.01))(bottleneck) \n",
    "output_layer = Dense(784, activation='sigmoid', kernel_regularizer=l2(0.01))(decoded) \n",
    "\n",
    "# Autoencoder model with L2 regularization \n",
    "autoencoder_regularized = Model(input_layer, output_layer) \n",
    "autoencoder_regularized.compile(optimizer='adam', loss='binary_crossentropy') \n",
    "\n",
    "# Train the model \n",
    "autoencoder_regularized.fit(x_train, x_train, epochs=50, batch_size=256, shuffle=True, validation_data=(x_test, x_test)) \n",
    "\n",
    "# Evaluate the model \n",
    "loss = autoencoder_regularized.evaluate(x_test, x_test) \n",
    "print(f'Regularized Autoencoder - Test loss: {loss}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "\n",
    "from tensorflow.keras.regularizers import l2 \n",
    "\n",
    "# Encoder with L2 regularization \n",
    "input_layer = Input(shape=(784,)) \n",
    "encoded = Dense(64, activation='relu', kernel_regularizer=l2(0.01))(input_layer) \n",
    "bottleneck = Dense(32, activation='relu', kernel_regularizer=l2(0.01))(encoded) \n",
    "\n",
    "# Decoder with L2 regularization \n",
    "decoded = Dense(64, activation='relu', kernel_regularizer=l2(0.01))(bottleneck) \n",
    "output_layer = Dense(784, activation='sigmoid', kernel_regularizer=l2(0.01))(decoded) \n",
    "\n",
    "# Autoencoder model with L2 regularization \n",
    "autoencoder_regularized = Model(input_layer, output_layer) \n",
    "autoencoder_regularized.compile(optimizer='adam', loss='binary_crossentropy') \n",
    "\n",
    "# Train the model \n",
    "autoencoder_regularized.fit(x_train, x_train, epochs=50, batch_size=256, shuffle=True, validation_data=(x_test, x_test)) \n",
    "\n",
    "# Evaluate the model \n",
    "loss = autoencoder_regularized.evaluate(x_test, x_test) \n",
    "print(f'Regularized Autoencoder - Test loss: {loss}') \n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 3 - Visualizing Intermediate Representations \n",
    "\n",
    "#### Objective: \n",
    "\n",
    "To visualize and understand the intermediate representations (encoded features) learned by the autoencoder. \n",
    "\n",
    "#### Instructions: \n",
    "\n",
    "**1. Extract Encoder Part:**\n",
    "- Extract the encoder part of the trained autoencoder to create a separate model that outputs the encoded features. \n",
    "\n",
    "**2. Visualize Encoded Features:**\n",
    "- Use the encoder model to transform the test data into the encoded space. \n",
    "- Plot the encoded features using a scatter plot for the first two dimensions of the encoded space. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2IAAAK9CAYAAABGsN/cAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABXU0lEQVR4nO3deVxV1f7/8fcBGRQFUhEkUcjMcErTJFKylCsOOZTd1CxR+WZ1My3L0ttgs1ODNmjpTSvTq5nNmbOVKWnOs5k5lAoOBCgmIKzfH/441xOgoIdF4Ov5eJxHnrXX3vuzl5uTb/bZazuMMUYAAAAAAGs8SrsAAAAAALjUEMQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAMCC9957Tw6HQ3v37nXbNvv166fw8HC3bQ8lg7+ni/ftt9/K4XDo22+/Le1SLlp4eLj69etX2mUA+BsgiAEoU/ICTWGvH3/8sbRL/Nu56aabCh2vHTt2lMg+J06cqPfee69Etv13Ex4eXuj4njp1yu37W7lypZ555hmlpqaes19eeCnKy6a9e/e67NvLy0vVq1fXDTfcoH//+9/av3+/1XoAoLRUKO0CAOBCPPfcc4qIiMjXfuWVV5ZCNX9/tWrV0qhRo/K1h4aGlsj+Jk6cqOrVq18yv/lv2rSpHnnkkXzt3t7emjJlinJzc922r5UrV+rZZ59Vv379FBgYWGi/yMhITZ8+3aVtxIgRqly5sp544gm31XOhevfurU6dOik3N1d//PGHfvrpJ40fP14TJkzQu+++q169ejn73njjjfrzzz/l7e1dihW7x86dO+Xhwe/BARDEAJRRHTt2VIsWLUq7jDIjICBAd911V2mXcVGMMTp16pQqVqxY2qXkc/nllxc6vkX5R/fp06eVm5vr1qARHBycr6bRo0erevXqf4tz4dprr81Xx759+9S+fXvFx8crMjJS11xzjaQzY+jr61saZbqdj49PaZcA4G+CX8kAKJfyvv708ssva/Lkyapbt658fHx03XXX6aeffsrXf8eOHbrjjjsUFBSkihUrqn79+vmuGqxfv14dO3aUv7+/KleurHbt2hX4VcitW7eqbdu2qlixomrVqqUXXnih0Csi33zzjWJiYuTn56cqVaqoc+fO2rp1a75+n332mRo1aiRfX181atRIn3766QWOTMEyMzM1cuRIXXnllfLx8VFYWJgee+wxZWZmuvSbNm2a2rZtqxo1asjHx0cNGjTQpEmTXPqEh4dr69at+u6775xfP7vpppskSc8880yBX4Ur6B668PBw3XLLLVqwYIFatGihihUr6p133pEkpaam6qGHHlJYWJh8fHx05ZVXasyYMfnGedasWWrevLmqVKkif39/NW7cWBMmTHDDiBXdX+8RO/vcHD9+vPPc3LZtmyTpjTfeUMOGDVWpUiVddtllatGihWbOnCnpzPgNGzZMkhQREeEc3wu599AYo+rVq2vo0KHOttzcXAUGBsrT09Plq49jxoxRhQoVdOLECWfb0qVLneduYGCgunXrpu3btxe7jrPVqVNH7733nrKysjR27Fhne0H3iN10001q1KiRNm3apDZt2qhSpUq68sor9fHHH0uSvvvuO0VFRTl/nhcvXpxvfwcOHNCAAQMUHBwsHx8fNWzYUFOnTnXpk7fvjz76SC+++KJq1aolX19ftWvXTr/88otL3127dqlHjx4KCQmRr6+vatWqpV69eiktLc3Zp6B7xH799Vf985//VNWqVVWpUiVdf/31+vrrr0u0DgCljytiAMqktLQ0HT161KXN4XCoWrVqLm0zZ87U8ePHde+998rhcGjs2LG67bbb9Ouvv8rLy0uStGnTJsXExMjLy0sDBw5UeHi4du/erS+//FIvvviipDPhKiYmRv7+/nrsscfk5eWld955RzfddJPzH3ySlJSUpJtvvlmnT5/W8OHD5efnp8mTJxd4FWf69OmKj49XXFycxowZo5MnT2rSpElq3bq11q9f7/zH+8KFC9WjRw81aNBAo0aN0rFjx9S/f3/VqlWryOOVk5OTb7x8fX1VuXJl5ebmqmvXrvrhhx80cOBARUZGavPmzXrttdf0888/67PPPnOuM2nSJDVs2FBdu3ZVhQoV9OWXX+pf//qXcnNz9cADD0iSxo8frwcffNDlK3DBwcFFrvVsO3fuVO/evXXvvffqnnvuUf369XXy5Em1adNGBw4c0L333qvatWtr5cqVGjFihA4dOqTx48dLkhYtWqTevXurXbt2GjNmjCRp+/btWrFihYYMGXJB9RQmOzs73/hWqlRJlSpVKnSdadOm6dSpUxo4cKB8fHxUtWpVTZkyRYMHD9btt9+uIUOG6NSpU9q0aZNWrVqlO++8U7fddpt+/vln/fe//9Vrr72m6tWrS5KCgoKKXbPD4VCrVq30/fffO9s2bdqktLQ0eXh4aMWKFercubMkafny5WrWrJkqV64sSVq8eLE6duyoK664Qs8884z+/PNPvfHGG2rVqpXWrVt3UZOTREdHq27dulq0aNF5+/7xxx+65ZZb1KtXL/3zn//UpEmT1KtXL82YMUMPPfSQ7rvvPt15550aN26cbr/9dv3222+qUqWKJCk5OVnXX3+9HA6HBg0apKCgIH3zzTdKSEhQenq6HnroIZd9jR49Wh4eHnr00UeVlpamsWPHqk+fPlq1apUkKSsrS3FxccrMzNSDDz6okJAQHThwQF999ZVSU1MVEBBQ4DEkJyfrhhtu0MmTJzV48GBVq1ZN77//vrp27aqPP/5Yt956q5U6AJQCAwBlyLRp04ykAl8+Pj7Ofnv27DGSTLVq1UxKSoqz/fPPPzeSzJdffulsu/HGG02VKlXMvn37XPaVm5vr/HP37t2Nt7e32b17t7Pt4MGDpkqVKubGG290tj300ENGklm1apWz7fDhwyYgIMBIMnv27DHGGHP8+HETGBho7rnnHpd9JiUlmYCAAJf2pk2bmpo1a5rU1FRn28KFC40kU6dOnfOOWZs2bQocr/j4eGOMMdOnTzceHh5m+fLlLuu9/fbbRpJZsWKFs+3kyZP5th8XF2euuOIKl7aGDRuaNm3a5Os7cuRIU9D/evL+XvPGxxhj6tSpYySZ+fPnu/R9/vnnjZ+fn/n5559d2ocPH248PT3N/v37jTHGDBkyxPj7+5vTp0/nHxQ3yqvzr6+RI0caY4yJj493+XvKOzf9/f3N4cOHXbbVrVs307Bhw3Pub9y4cfnGqqj++vcybtw44+npadLT040xxrz++uumTp06pmXLlubxxx83xhiTk5NjAgMDzcMPP+xcr2nTpqZGjRrm2LFjzraNGzcaDw8P07dv33PWkHf848aNK7RPt27djCSTlpZmjDFm2bJlRpJZtmyZs0/eeT1z5kxn244dO4wk4+HhYX788Udn+4IFC4wkM23aNGdbQkKCqVmzpjl69KjLvnv16mUCAgKc53reviMjI01mZqaz34QJE4wks3nzZmOMMevXrzeSzJw5c855/HXq1HH+7Bnzv8+Ms3/+jh8/biIiIkx4eLjJyckpkToAlD6+mgigTHrrrbe0aNEil9c333yTr1/Pnj112WWXOd/HxMRIOvNVIEk6cuSIvv/+ew0YMEC1a9d2WTfvK3Q5OTlauHChunfvriuuuMK5vGbNmrrzzjv1ww8/KD09XZI0b948XX/99WrZsqWzX1BQkPr06eOy7UWLFik1NVW9e/fW0aNHnS9PT09FRUVp2bJlkqRDhw5pw4YNio+Pd/lN9j/+8Q81aNCgyOMVHh6eb7wee+wxSdKcOXMUGRmpq6++2qWWtm3bSpKzFkkuV/byrkq2adNGv/76a4l87SkiIkJxcXEubXPmzFFMTIwuu+wyl3pjY2OVk5PjvMITGBiojIyMIl1ZuVhRUVH5xrdv377nXKdHjx75rmQFBgbq999/L/DrsyUhJiZGOTk5WrlypaQzV75iYmIUExOj5cuXS5K2bNmi1NRU589O3jnZr18/Va1a1bmtJk2a6B//+IfmzZt30XXlXXk7fvz4efudPalH/fr1FRgYqMjISOdVaknOP+f93BtjNHfuXHXp0kXGGJfzKC4uTmlpaVq3bp3Lvvr37+9yD99fP0vyfj4XLFigkydPFvlY582bp5YtW6p169YuxzVw4EDt3bvX+ZXVkq4DgH18NRFAmdSyZcsiTdbx13CVF8r++OMPSf/7x0ujRo0K3caRI0d08uRJ1a9fP9+yyMhI5ebm6rffflPDhg21b98+l38A5vnrurt27ZIkZ9j5K39/f0lnJi+QpHr16hW4zb/+Y7Ewfn5+io2NLXDZrl27tH379kK/3nb48GHnn1esWKGRI0cqMTEx3z/y0tLS3P61p4Jmxty1a5c2bdp03nr/9a9/6aOPPlLHjh11+eWXq3379rrjjjvUoUOHc+7zyJEjysnJcb6vXLmyMxgUpnr16oWOb2EKOrbHH39cixcvVsuWLXXllVeqffv2uvPOO9WqVatibbuorr32WlWqVEnLly9XXFycli9frmeffVYhISF64403dOrUKWcgywsKeedkYT8PCxYsUEZGhvz8/C64rrx70fK+RliYWrVq5bvnMCAgQGFhYfnapP/93B85ckSpqamaPHmyJk+eXOC2zz7vpfN/lkRERGjo0KF69dVXNWPGDMXExKhr16666667zvlzUdhnRmRkpHP52Z9PJVUHAPsIYgDKNU9PzwLbjTGWK3GVN6nE9OnTFRISkm95hQr2Pp5zc3PVuHFjvfrqqwUuz/tH7e7du9WuXTtdffXVevXVVxUWFiZvb2/NmzdPr732WpGmaC/smVVnB5+zFXRvXW5urv7xj384r+j91VVXXSVJqlGjhjZs2KAFCxbom2++0TfffKNp06apb9++ev/99wut8brrrnOGDUkaOXKknnnmmUL7X6iCji0yMlI7d+7UV199pfnz52vu3LmaOHGinn76aT377LNur8HLy0tRUVH6/vvv9csvvygpKUkxMTEKDg5Wdna2Vq1apeXLl+vqq6++oPvQLtSWLVtUo0YN5y8kClPYz/f5fu7zztW77rpL8fHxBfZt0qRJsbYpSa+88or69eunzz//XAsXLtTgwYM1atQo/fjjj8W6p/Nc/i51ALh4BDEAl7S8rxpu2bKl0D5BQUGqVKmSdu7cmW/Zjh075OHh4QwrderUcV7tOttf161bt66kM2HhXFdS6tSpI0lF2uaFqlu3rjZu3Kh27dqd8+G+X375pTIzM/XFF1+4/Fb+7K8u5ilsO3m/vU9NTXV5BtbZwaco9Z44caJIV6C8vb3VpUsXdenSRbm5ufrXv/6ld955R0899VShz5ybMWOG/vzzT+f7s7+OaoOfn5969uypnj17KisrS7fddptefPFFjRgxQr6+vm5/AHNMTIzGjBmjxYsXq3r16rr66qvlcDjUsGFDLV++XMuXL9ctt9zi7J93Thb281C9evWLuhqWmJio3bt3l+gU+0FBQapSpYpycnKKfSXzfBo3bqzGjRvrySef1MqVK9WqVSu9/fbbeuGFFwrsX6dOnULHMm+5jToA2Mc9YgAuaUFBQbrxxhs1depU7d+/32VZ3m+YPT091b59e33++ecu04QnJydr5syZat26tfM39506ddKPP/6o1atXO/sdOXJEM2bMcNl2XFyc/P399dJLLyk7OztfXUeOHJF05j60pk2b6v3333e5B2vRokX57h25UHfccYcOHDigKVOm5Fv2559/KiMjQ9L/fhN/9m/e09LSNG3atHzr+fn5uUx/nicvgJ49U19GRsY5r1AVVG9iYqIWLFiQb1lqaqpOnz4tSTp27JjLMg8PD+dVjr9Oy3+2Vq1aKTY21vmyGcT+WrO3t7caNGggY4zzPMkLOQWN74WIiYlRZmamxo8fr9atWzuDXkxMjKZPn66DBw8670OSXM/Js2vYsmWLFi5cqE6dOl1wLfv27VO/fv3k7e3tnKa/JHh6eqpHjx6aO3dugb+Eyfv5K4709HTnuZencePG8vDwOOf51qlTJ61evVqJiYnOtoyMDE2ePFnh4eHFuhf0YuoAYB9XxACUSd98843zN8Znu+GGG4r9D+fXX39drVu31rXXXquBAwcqIiJCe/fu1ddff60NGzZIkl544QUtWrRIrVu31r/+9S9VqFBB77zzjjIzM12ed/TYY49p+vTp6tChg4YMGeKcvr5OnTratGmTs5+/v78mTZqku+++W9dee6169eqloKAg7d+/X19//bVatWqlN998U5I0atQode7cWa1bt9aAAQOUkpLifNbU2c91ulB33323PvroI913331atmyZWrVqpZycHO3YsUMfffSR8zle7du3d15huvfee3XixAlNmTJFNWrU0KFDh1y22bx5c02aNEkvvPCCrrzyStWoUUNt27ZV+/btVbt2bSUkJGjYsGHy9PTU1KlTncdeFMOGDdMXX3yhW265Rf369VPz5s2VkZGhzZs36+OPP9bevXtVvXp1/d///Z9SUlLUtm1b1apVS/v27dMbb7yhpk2bOu+/+btp3769QkJC1KpVKwUHB2v79u1688031blzZ+f9Us2bN5ckPfHEE+rVq5e8vLzUpUuXC74KFR0drQoVKmjnzp0aOHCgs/3GG290PiPu7CAmSePGjVPHjh0VHR2thIQE5/T1AQEBRf4a57p16/Thhx8qNzdXqamp+umnnzR37lw5HA5Nnz4931cD3W306NFatmyZoqKidM8996hBgwZKSUnRunXrtHjxYqWkpBRre0uXLtWgQYP0z3/+U1dddZVOnz6t6dOnO0NfYYYPH67//ve/6tixowYPHqyqVavq/fff1549ezR37twiPRDcHXUAKAWlNl8jAFyAc01fr7Ompz7XFNk6a2rxPFu2bDG33nqrCQwMNL6+vqZ+/frmqaeecumzbt06ExcXZypXrmwqVapkbr75ZrNy5cp829+0aZNp06aN8fX1NZdffrl5/vnnzbvvvlvglOPLli0zcXFxJiAgwPj6+pq6deuafv36mTVr1rj0mzt3romMjDQ+Pj6mQYMG5pNPPsk3LXph2rRpc94p0bOyssyYMWNMw4YNjY+Pj7nssstM8+bNzbPPPuucQtwYY7744gvTpEkT4+vra8LDw82YMWPM1KlT8x1bUlKS6dy5s6lSpYqR5DJl+tq1a01UVJTx9vY2tWvXNq+++mqh09d37ty5wHqPHz9uRowYYa688krj7e1tqlevbm644Qbz8ssvm6ysLGOMMR9//LFp3769qVGjhnNf9957rzl06NB5x6w4zlWnMYVPX1/QufnOO++YG2+80VSrVs34+PiYunXrmmHDhrn8HRhzZgr/yy+/3Hh4eBRrKvvCHitw3XXX5Xvswu+//24kmbCwsAK3tXjxYtOqVStTsWJF4+/vb7p06WK2bdt23hryjj/vVaFCBVO1alUTFRVlRowYke8xEsYUPn19Qed1YX8fkswDDzzg0pacnGweeOABExYWZry8vExISIhp166dmTx5cr59/3U6+LzjyPvM+fXXX82AAQNM3bp1ja+vr6lataq5+eabzeLFi/PVd/b09cYYs3v3bnP77bc7P39atmxpvvrqqwLHwF11ACh9DmNK+Y51AAAAALjEcI8YAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwHOrtBbm6uDh48qCpVqsjhcJR2OQAAAABKiTFGx48fV2ho6Dkfyk4Qc4ODBw8qLCystMsAAAAA8Dfx22+/qVatWoUuJ4i5QZUqVSSdGWx/f/9SrgYAAABAaUlPT1dYWJgzIxSGIOYGeV9H9Pf3J4gBAAAAOO8tS0zWAQAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgWZkLYm+99ZbCw8Pl6+urqKgorV69+pz958yZo6uvvlq+vr5q3Lix5s2bV2jf++67Tw6HQ+PHj3dz1QAAAADwP2UqiM2ePVtDhw7VyJEjtW7dOl1zzTWKi4vT4cOHC+y/cuVK9e7dWwkJCVq/fr26d++u7t27a8uWLfn6fvrpp/rxxx8VGhpa0ocBAAAA4BJXpoLYq6++qnvuuUf9+/dXgwYN9Pbbb6tSpUqaOnVqgf0nTJigDh06aNiwYYqMjNTzzz+va6+9Vm+++aZLvwMHDujBBx/UjBkz5OXlZeNQAAAAAFzCykwQy8rK0tq1axUbG+ts8/DwUGxsrBITEwtcJzEx0aW/JMXFxbn0z83N1d13361hw4apYcOGRaolMzNT6enpLi8AAAAAKKoyE8SOHj2qnJwcBQcHu7QHBwcrKSmpwHWSkpLO23/MmDGqUKGCBg8eXORaRo0apYCAAOcrLCysGEcCAAAA4FJXZoJYSVi7dq0mTJig9957Tw6Ho8jrjRgxQmlpac7Xb7/9VoJVAgAAAChvykwQq169ujw9PZWcnOzSnpycrJCQkALXCQkJOWf/5cuX6/Dhw6pdu7YqVKigChUqaN++fXrkkUcUHh5eaC0+Pj7y9/d3eQEAAABAUZWZIObt7a3mzZtryZIlzrbc3FwtWbJE0dHRBa4THR3t0l+SFi1a5Ox/9913a9OmTdqwYYPzFRoaqmHDhmnBggUldzAAAAAALmkVSruA4hg6dKji4+PVokULtWzZUuPHj1dGRob69+8vSerbt68uv/xyjRo1SpI0ZMgQtWnTRq+88oo6d+6sWbNmac2aNZo8ebIkqVq1aqpWrZrLPry8vBQSEqL69evbPTgAAAAAl4wyFcR69uypI0eO6Omnn1ZSUpKaNm2q+fPnOyfk2L9/vzw8/neR74YbbtDMmTP15JNP6t///rfq1aunzz77TI0aNSqtQwAAAAAAOYwxprSLKOvS09MVEBCgtLQ07hcDAAAALmFFzQZl5h4xAAAAACgvCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwrMwFsbfeekvh4eHy9fVVVFSUVq9efc7+c+bM0dVXXy1fX181btxY8+bNcy7Lzs7W448/rsaNG8vPz0+hoaHq27evDh48WNKHAQAAAOASVqaC2OzZszV06FCNHDlS69at0zXXXKO4uDgdPny4wP4rV65U7969lZCQoPXr16t79+7q3r27tmzZIkk6efKk1q1bp6eeekrr1q3TJ598op07d6pr1642DwsAAADAJcZhjDGlXURRRUVF6brrrtObb74pScrNzVVYWJgefPBBDR8+PF//nj17KiMjQ1999ZWz7frrr1fTpk319ttvF7iPn376SS1bttS+fftUu3btItWVnp6ugIAApaWlyd/f/wKODAAAAEB5UNRsUGauiGVlZWnt2rWKjY11tnl4eCg2NlaJiYkFrpOYmOjSX5Li4uIK7S9JaWlpcjgcCgwMLLRPZmam0tPTXV4AAAAAUFRlJogdPXpUOTk5Cg4OdmkPDg5WUlJSgeskJSUVq/+pU6f0+OOPq3fv3udMr6NGjVJAQIDzFRYWVsyjAQAAAHApKzNBrKRlZ2frjjvukDFGkyZNOmffESNGKC0tzfn67bffLFUJAAAAoDyoUNoFFFX16tXl6emp5ORkl/bk5GSFhIQUuE5ISEiR+ueFsH379mnp0qXnvc/Lx8dHPj4+F3AUAAAAAFCGroh5e3urefPmWrJkibMtNzdXS5YsUXR0dIHrREdHu/SXpEWLFrn0zwthu3bt0uLFi1WtWrWSOQAAAAAA+P/KzBUxSRo6dKji4+PVokULtWzZUuPHj1dGRob69+8vSerbt68uv/xyjRo1SpI0ZMgQtWnTRq+88oo6d+6sWbNmac2aNZo8ebKkMyHs9ttv17p16/TVV18pJyfHef9Y1apV5e3tXToHCgAAAKBcK1NBrGfPnjpy5IiefvppJSUlqWnTppo/f75zQo79+/fLw+N/F/luuOEGzZw5U08++aT+/e9/q169evrss8/UqFEjSdKBAwf0xRdfSJKaNm3qsq9ly5bppptusnJcAAAAAC4tZeo5Yn9XPEcMAAAAgFQOnyMGAAAAAOUFQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsK1YQ27hxo1544QVNnDhRR48edVmWnp6uAQMGuLU4AAAAACiPHMYYU5SOCxcuVJcuXVSvXj0dP35cGRkZmjNnjm6++WZJUnJyskJDQ5WTk1OiBf8dpaenKyAgQGlpafL39y/tcgAAAACUkqJmgyJfEXvmmWf06KOPasuWLdq7d68ee+wxde3aVfPnz3dLwQAAAABwqahQ1I5bt27V9OnTJUkOh0OPPfaYatWqpdtvv12zZs3SddddV2JFAgAAAEB5UuQg5uPjo9TUVJe2O++8Ux4eHurZs6deeeUVd9cGAAAAAOVSkYNY06ZNtWzZMjVv3tylvVevXjLGKD4+3u3FAQAAAEB5VOQgdv/99+v7778vcFnv3r1ljNGUKVPcVhgAAAAAlFdFnjURhWPWRAAAAABSCcyaCAAAAABwD4IYAAAAAFhGEAMAAAAAywhiAAAAAGDZBQexrKws7dy5U6dPn3ZnPQAAAABQ7hU7iJ08eVIJCQmqVKmSGjZsqP3790uSHnzwQY0ePdrtBQIAAABAeVPsIDZixAht3LhR3377rXx9fZ3tsbGxmj17tluLAwAAAIDyqMgPdM7z2Wefafbs2br++uvlcDic7Q0bNtTu3bvdWhwAAAAAlEfFviJ25MgR1ahRI197RkaGSzADAAAAABSs2EGsRYsW+vrrr53v88LXf/7zH0VHR7uvMgAAAAAop4r91cSXXnpJHTt21LZt23T69GlNmDBB27Zt08qVK/Xdd9+VRI0AAAAAUK4U+4pY69attXHjRp0+fVqNGzfWwoULVaNGDSUmJqp58+YlUSMAAAAAlCvFuiKWnZ2te++9V0899ZSmTJlSUjUBAAAAQLlWrCtiXl5emjt3bknVAgAAAACXhGJ/NbF79+767LPPSqAUAAAAALg0FHuyjnr16um5557TihUr1Lx5c/n5+bksHzx4sNuKAwAAAIDyyGGMMcVZISIiovCNORz69ddfL7qosiY9PV0BAQFKS0uTv79/aZcDAAAAoJQUNRsU+4rYnj17LqowAAAAALjUFfseMQAAAADAxSn2FbEBAwacc/nUqVMvuBgAAAAAuBQUO4j98ccfLu+zs7O1ZcsWpaamqm3btm4rDAAAAADKq2IHsU8//TRfW25uru6//37VrVvXLUUBAAAAQHnmlnvEPDw8NHToUL322mvu2BwAAAAAlGtum6xj9+7dOn36tLs2BwAAAADlVrG/mjh06FCX98YYHTp0SF9//bXi4+PdVhgAAAAAlFfFDmLr1693ee/h4aGgoCC98sor551REQAAAABwAUFs2bJlJVEHAAAAAFwyin2PWNu2bZWampqvPT09nenrAQAAAKAIih3Evv32W2VlZeVrP3XqlJYvX+6WogAAAACgPCvyVxM3bdrk/PO2bduUlJTkfJ+Tk6P58+fr8ssvd291AAAAAFAOFTmINW3aVA6HQw6Ho8CvIFasWFFvvPGGW4sDAAAAgPKoyEFsz549Msboiiuu0OrVqxUUFORc5u3trRo1asjT07NEigQAAACA8qTIQaxOnTqSpNzc3BIrBgAAAAAuBcWevj7Ptm3btH///nwTd3Tt2vWiiwIAAACA8qzYQezXX3/Vrbfeqs2bN8vhcMgYI0lyOBySzkzcAQAAAAAoXLGnrx8yZIgiIiJ0+PBhVapUSVu3btX333+vFi1a6Ntvvy2BEgEAAACgfCn2FbHExEQtXbpU1atXl4eHhzw8PNS6dWuNGjVKgwcP1vr160uiTgAAAAAoN4p9RSwnJ0dVqlSRJFWvXl0HDx6UdGYyj507d7q3OgAAAAAoh4p9RaxRo0bauHGjIiIiFBUVpbFjx8rb21uTJ0/WFVdcURI1AgAAAEC5Uuwg9uSTTyojI0OS9Nxzz+mWW25RTEyMqlWrptmzZ7u9QAAAAAAobxwmb9rDi5CSkqLLLrvMOXPipSY9PV0BAQFKS0uTv79/aZcDAAAAoJQUNRsU+x6xPL/88osWLFigP//8U1WrVr3QzQAAAADAJafYQezYsWNq166drrrqKnXq1EmHDh2SJCUkJOiRRx5xe4EAAAAAUN4UO4g9/PDD8vLy0v79+1WpUiVne8+ePTV//ny3FgcAAAAA5VGxJ+tYuHChFixYoFq1arm016tXT/v27XNbYQAAAABQXhX7ilhGRobLlbA8KSkp8vHxcUtRAAAAAFCeFTuIxcTE6IMPPnC+dzgcys3N1dixY3XzzTe7tTgAAAAAKI+K/dXEsWPHql27dlqzZo2ysrL02GOPaevWrUpJSdGKFStKokYAAAAAKFeKfUWsUaNG+vnnn9WqVSt169ZNGRkZuu2227R+/XrVrVu3JGoEAAAAgHKlyA90njp1qvr06cN9YAXggc4AAAAApBJ4oPM999yjtLQ05/vQ0FDt3bv3oooEAKA8cDjyvwAAOJciB7G/Xjg7fvy4cnNz3V7Q+bz11lsKDw+Xr6+voqKitHr16nP2nzNnjq6++mr5+vqqcePGmjdvnstyY4yefvpp1axZUxUrVlRsbKx27dpVkocAAChHCgtdhDEAwLkU+x6x0jR79mwNHTpUI0eO1Lp163TNNdcoLi5Ohw8fLrD/ypUr1bt3byUkJGj9+vXq3r27unfvri1btjj7jB07Vq+//rrefvttrVq1Sn5+foqLi9OpU6dsHRYAoIw6X9gijAEAClPke8Q8PT2VlJSkoKAgSZK/v782btyoiIiIEi3wbFFRUbruuuv05ptvSpJyc3MVFhamBx98UMOHD8/Xv2fPnsrIyNBXX33lbLv++uvVtGlTvf322zLGKDQ0VI888ogeffRRSVJaWpqCg4P13nvvqVevXkWqi3vEAODSU5yQVbT/0wIAygO33yNmjNFVV12lqlWrqmrVqjpx4oSaNWvmfJ/3KilZWVlau3atYmNjnW0eHh6KjY1VYmJigeskJia69JekuLg4Z/89e/YoKSnJpU9AQICioqIK3aYkZWZmKj093eUFAAAAAEVV5OeITZs2rSTrOK+jR48qJydHwcHBLu3BwcHasWNHgeskJSUV2D8pKcm5PK+tsD4FGTVqlJ599tliHwMAAAAASMUIYvHx8SVZR5kyYsQIDR061Pk+PT1dYWFhpVgRAAAAgLKkzEzWUb16dXl6eio5OdmlPTk5WSEhIQWuExIScs7+ef8tzjYlycfHR/7+/i4vAAAAACiqMhPEvL291bx5cy1ZssTZlpubqyVLlig6OrrAdaKjo136S9KiRYuc/SMiIhQSEuLSJz09XatWrSp0mwAASEWfgIOJOgAABSnyVxP/DoYOHar4+Hi1aNFCLVu21Pjx45WRkaH+/ftLkvr27avLL79co0aNkiQNGTJEbdq00SuvvKLOnTtr1qxZWrNmjSZPnixJcjgceuihh/TCCy+oXr16ioiI0FNPPaXQ0FB17969tA4TAFBGGHPu2RMJYQCAwpSpINazZ08dOXJETz/9tJKSktS0aVPNnz/fOdnG/v375eHxv4t8N9xwg2bOnKknn3xS//73v1WvXj199tlnatSokbPPY489poyMDA0cOFCpqalq3bq15s+fL19fX+vHBwAoewoLY4QwAMC5FPk5YigczxEDAAAAIBU9GxTpitjZMwSez6uvvlrkvgAAAABwKSpSEFu/fr3L+3Xr1un06dOqX7++JOnnn3+Wp6enmjdv7v4KAQAAAKCcKVIQW7ZsmfPPr776qqpUqaL3339fl112mSTpjz/+UP/+/RUTE1MyVQIAAABAOVLse8Quv/xyLVy4UA0bNnRp37Jli9q3b6+DBw+6tcCygHvEAAAAAEhFzwbFfo5Yenq6jhw5kq/9yJEjOn78eHE3BwAAAACXnGIHsVtvvVX9+/fXJ598ot9//12///675s6dq4SEBN12220lUSMAAAAAlCvFfo7Y22+/rUcffVR33nmnsrOzz2ykQgUlJCRo3Lhxbi8QAAAAAMqbC36OWEZGhnbv3i1Jqlu3rvz8/NxaWFnCPWIAAAAApBK8RyzPoUOHdOjQIdWrV09+fn7iudAAAAAAUDTFDmLHjh1Tu3btdNVVV6lTp046dOiQJCkhIUGPPPKI2wsEAAAAgPKm2EHs4YcflpeXl/bv369KlSo523v27Kn58+e7tTgAAAAAKI+KPVnHwoULtWDBAtWqVculvV69etq3b5/bCgMAAACA8qrYV8QyMjJcroTlSUlJkY+Pj1uKAgAAAIDyrNhBLCYmRh988IHzvcPhUG5ursaOHaubb77ZrcUBAAAAQHlU7K8mjh07Vu3atdOaNWuUlZWlxx57TFu3blVKSopWrFhREjUCAAAAQLlS7CtijRo10s8//6zWrVurW7duysjI0G233ab169erbt26JVEjAAAAAJQrF/xAZ/wPD3QGAAAAIBU9GxTpq4mbNm0q8o6bNGlS5L4AAAAAcCkqUhBr2rSpHA6HjDFyOBzO9ryLaWe35eTkuLlEAAAAAChfinSP2J49e/Trr79qz549mjt3riIiIjRx4kRt2LBBGzZs0MSJE1W3bl3NnTu3pOsFAAAAgDKvSFfE6tSp4/zzP//5T73++uvq1KmTs61JkyYKCwvTU089pe7du7u9SAAAAAAoT4o9a+LmzZsVERGRrz0iIkLbtm1zS1EAAAAAUJ4VO4hFRkZq1KhRysrKcrZlZWVp1KhRioyMdGtxAAAAAFAeFfuBzm+//ba6dOmiWrVqOWdI3LRpkxwOh7788ku3FwgAAAAA5c0FPUcsIyNDM2bM0I4dOySduUp25513ys/Pz+0FlgU8RwwAAACA5ObniP2Vn5+fBg4ceMHFAQAAAMCl7IKC2O7duzV+/Hht375dktSwYUMNHjxYdevWdWtxAAAAAFAeFXuyjgULFqhBgwZavXq1mjRpoiZNmujHH39Uw4YNtWjRopKoEQAAAADKlWLfI9asWTPFxcVp9OjRLu3Dhw/XwoULtW7dOrcWWBZwjxgAAAAAqejZoNhXxLZv366EhIR87QMGDOA5YgAAAABQBMUOYkFBQdqwYUO+9g0bNqhGjRruqAkAAAAAyrViT9Zxzz33aODAgfr11191ww03SJJWrFihMWPGaOjQoW4vEAAAAADKm2LfI2aM0fjx4/XKK6/o4MGDkqTQ0FANGzZMgwcPlsPhKJFC/864RwwAAACAVPRscEEPdM5z/PhxSVKVKlUudBPlAkEMAAAAgFSCD3Tes2ePTp8+rXr16rkEsF27dsnLy0vh4eEXVDAAAAAAXCqKPVlHv379tHLlynztq1atUr9+/dxREwAAAACUa8UOYuvXr1erVq3ytV9//fUFzqYIAAAAAHBV7CDmcDic94adLS0tTTk5OW4pCgAAAADKs2IHsRtvvFGjRo1yCV05OTkaNWqUWrdu7dbiAAAAAKA8KvZkHWPGjNGNN96o+vXrKyYmRpK0fPlypaena+nSpW4vEAAAAADKm2JfEWvQoIE2bdqkO+64Q4cPH9bx48fVt29f7dixQ40aNSqJGgEAAACgXLmo54jhDJ4jBgAAAEAqweeISVJqaqpWr16tw4cPKzc312VZ3759L2STAAAAAHDJKHYQ+/LLL9WnTx+dOHFC/v7+cjgczmUOh4MgBgAAAADnUex7xB555BENGDBAJ06cUGpqqv744w/nKyUlpSRqBAAAAIBypdhB7MCBAxo8eLAqVapUEvUAAAAAQLlX7CAWFxenNWvWlEQtAAAAAHBJKPY9Yp07d9awYcO0bds2NW7cWF5eXi7Lu3bt6rbiAAAAAKA8Kvb09R4ehV9EczgcysnJueiiyhqmrwcAAAAgleD09X+drh4AAAAAUDzFvkcMAAAAAHBxihzEOnXqpLS0NOf70aNHKzU11fn+2LFjatCggVuLAwAAAIDyqMhBbMGCBcrMzHS+f+mll1yeG3b69Gnt3LnTvdUBAAAAQDlU5CD21zk9ijnHBwAAAADg/+MeMQAAAACwrMhBzOFwyOFw5GsDAAAAABRPkaevN8aoX79+8vHxkSSdOnVK9913n/z8/CTJ5f4xAAAAAEDhihzE4uPjXd7fdddd+fr07dv34isCAAAAgHKuyEFs2rRpJVkHAAAAAFwymKwDAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlZSaIpaSkqE+fPvL391dgYKASEhJ04sSJc65z6tQpPfDAA6pWrZoqV66sHj16KDk52bl848aN6t27t8LCwlSxYkVFRkZqwoQJJX0oAAAAAC5xZSaI9enTR1u3btWiRYv01Vdf6fvvv9fAgQPPuc7DDz+sL7/8UnPmzNF3332ngwcP6rbbbnMuX7t2rWrUqKEPP/xQW7du1RNPPKERI0bozTffLOnDAQAAAHAJcxhjTGkXcT7bt29XgwYN9NNPP6lFixaSpPnz56tTp076/fffFRoamm+dtLQ0BQUFaebMmbr99tslSTt27FBkZKQSExN1/fXXF7ivBx54QNu3b9fSpUuLXF96eroCAgKUlpYmf3//CzhCAAAAAOVBUbNBmbgilpiYqMDAQGcIk6TY2Fh5eHho1apVBa6zdu1aZWdnKzY21tl29dVXq3bt2kpMTCx0X2lpaapateo568nMzFR6errLCwAAAACKqkwEsaSkJNWoUcOlrUKFCqpataqSkpIKXcfb21uBgYEu7cHBwYWus3LlSs2ePfu8X3kcNWqUAgICnK+wsLCiHwwAAACAS16pBrHhw4fL4XCc87Vjxw4rtWzZskXdunXTyJEj1b59+3P2HTFihNLS0pyv3377zUqNAAAAAMqHCqW580ceeUT9+vU7Z58rrrhCISEhOnz4sEv76dOnlZKSopCQkALXCwkJUVZWllJTU12uiiUnJ+dbZ9u2bWrXrp0GDhyoJ5988rx1+/j4yMfH57z9AAAAAKAgpRrEgoKCFBQUdN5+0dHRSk1N1dq1a9W8eXNJ0tKlS5Wbm6uoqKgC12nevLm8vLy0ZMkS9ejRQ5K0c+dO7d+/X9HR0c5+W7duVdu2bRUfH68XX3zRDUcFAAAAAOdWJmZNlKSOHTsqOTlZb7/9trKzs9W/f3+1aNFCM2fOlCQdOHBA7dq10wcffKCWLVtKku6//37NmzdP7733nvz9/fXggw9KOnMvmHTm64ht27ZVXFycxo0b59yXp6dnkQJiHmZNBAAAACAVPRuU6hWx4pgxY4YGDRqkdu3aycPDQz169NDrr7/uXJ6dna2dO3fq5MmTzrbXXnvN2TczM1NxcXGaOHGic/nHH3+sI0eO6MMPP9SHH37obK9Tp4727t1r5bgAAAAAXHrKzBWxvzOuiAEAAACQytlzxAAAAACgPCGIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwLIyE8RSUlLUp08f+fv7KzAwUAkJCTpx4sQ51zl16pQeeOABVatWTZUrV1aPHj2UnJxcYN9jx46pVq1acjgcSk1NLYEjAAAAAIAzykwQ69Onj7Zu3apFixbpq6++0vfff6+BAweec52HH35YX375pebMmaPvvvtOBw8e1G233VZg34SEBDVp0qQkSgcAAAAAFw5jjCntIs5n+/btatCggX766Se1aNFCkjR//nx16tRJv//+u0JDQ/Otk5aWpqCgIM2cOVO33367JGnHjh2KjIxUYmKirr/+emffSZMmafbs2Xr66afVrl07/fHHHwoMDCxyfenp6QoICFBaWpr8/f0v7mABAAAAlFlFzQZl4opYYmKiAgMDnSFMkmJjY+Xh4aFVq1YVuM7atWuVnZ2t2NhYZ9vVV1+t2rVrKzEx0dm2bds2Pffcc/rggw/k4VG04cjMzFR6errLCwAAAACKqkwEsaSkJNWoUcOlrUKFCqpataqSkpIKXcfb2zvfla3g4GDnOpmZmerdu7fGjRun2rVrF7meUaNGKSAgwPkKCwsr3gEBAAAAuKSVahAbPny4HA7HOV87duwosf2PGDFCkZGRuuuuu4q9XlpamvP122+/lVCFAAAAAMqjCqW580ceeUT9+vU7Z58rrrhCISEhOnz4sEv76dOnlZKSopCQkALXCwkJUVZWllJTU12uiiUnJzvXWbp0qTZv3qyPP/5YkpR3u1z16tX1xBNP6Nlnny1w2z4+PvLx8SnKIQIAAABAPqUaxIKCghQUFHTeftHR0UpNTdXatWvVvHlzSWdCVG5urqKiogpcp3nz5vLy8tKSJUvUo0cPSdLOnTu1f/9+RUdHS5Lmzp2rP//807nOTz/9pAEDBmj58uWqW7fuxR4eAAAAABSoVINYUUVGRqpDhw6655579Pbbbys7O1uDBg1Sr169nDMmHjhwQO3atdMHH3ygli1bKiAgQAkJCRo6dKiqVq0qf39/Pfjgg4qOjnbOmPjXsHX06FHn/oozayIAAAAAFEeZCGKSNGPGDA0aNEjt2rWTh4eHevTooddff925PDs7Wzt37tTJkyedba+99pqzb2ZmpuLi4jRx4sTSKB8AAAAAnMrEc8T+7niOGAAAAACpnD1HDAAAAADKE4IYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAAAAAlhHEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBDAAAAAAsI4gBAAAAgGUEMQAAAACwjCAGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGBZhdIuoDwwxkiS0tPTS7kSAAAAAKUpLxPkZYTCEMTc4Pjx45KksLCwUq4EAAAAwN/B8ePHFRAQUOhyhzlfVMN55ebm6uDBg6pSpYocDkdpl+M26enpCgsL02+//SZ/f//SLqdcY6ztYJztYaztYaztYJztYaztYaxLhjFGx48fV2hoqDw8Cr8TjCtibuDh4aFatWqVdhklxt/fnx9OSxhrOxhnexhrexhrOxhnexhrexhr9zvXlbA8TNYBAAAAAJYRxAAAAADAMoIYCuXj46ORI0fKx8entEsp9xhrOxhnexhrexhrOxhnexhrexjr0sVkHQAAAABgGVfEAAAAAMAyghgAAAAAWEYQAwAAAADLCGIAAAAAYBlBrJz6/vvv1aVLF4WGhsrhcOizzz477zozZszQNddco0qVKqlmzZoaMGCAjh075lz+3nvvyeFwuLx8fX1dtmGM0dNPP62aNWuqYsWKio2N1a5du9x9eH8bJTHON910U75xdjgc6ty5s7NPv3798i3v0KFDSRzi38aFjPVbb72lyMhIVaxYUfXr19cHH3yQr8+cOXN09dVXy9fXV40bN9a8efNcll9q57RUMmM9ZcoUxcTE6LLLLtNll12m2NhYrV692qUP57V7xprP6vxKYpz5rC7YqFGjdN1116lKlSqqUaOGunfvrp07d553PXd8FqekpKhPnz7y9/dXYGCgEhISdOLECbce399FSYxzdna2Hn/8cTVu3Fh+fn4KDQ1V3759dfDgQZdthIeH5zuvR48e7fZjvCQYlEvz5s0zTzzxhPnkk0+MJPPpp5+es/8PP/xgPDw8zIQJE8yvv/5qli9fbho2bGhuvfVWZ59p06YZf39/c+jQIecrKSnJZTujR482AQEB5rPPPjMbN240Xbt2NREREebPP/8sicMsdSUxzseOHXMZ4y1bthhPT08zbdo0Z5/4+HjToUMHl34pKSkldJR/D8Ud64kTJ5oqVaqYWbNmmd27d5v//ve/pnLlyuaLL75w9lmxYoXx9PQ0Y8eONdu2bTNPPvmk8fLyMps3b3b2udTOaWNKZqzvvPNO89Zbb5n169eb7du3m379+pmAgADz+++/O/twXrtnrPmszq8kxpnP6oLFxcWZadOmmS1btpgNGzaYTp06mdq1a5sTJ04Uuo67Pos7dOhgrrnmGvPjjz+a5cuXmyuvvNL07t27RI+3tJTEOKempprY2Fgze/Zss2PHDpOYmGhatmxpmjdv7rKdOnXqmOeee87lvD7XflE4gtgloCj/0xk3bpy54oorXNpef/11c/nllzvfT5s2zQQEBBS6jdzcXBMSEmLGjRvnbEtNTTU+Pj7mv//97wXVXpa4a5z/6rXXXjNVqlRx+ZCLj4833bp1u5hyy7SijHV0dLR59NFHXdqGDh1qWrVq5Xx/xx13mM6dO7v0iYqKMvfee68xhnPaGPeN9V+dPn3aVKlSxbz//vvONs5r94w1n9XnVlLnNJ/VBTt8+LCRZL777rtC+7jjs3jbtm1Gkvnpp5+cfb755hvjcDjMgQMH3HlIf0vuGOeCrF692kgy+/btc7bVqVPHvPbaaxddM4zhq4mQJEVHR+u3337TvHnzZIxRcnKyPv74Y3Xq1Mml34kTJ1SnTh2FhYWpW7du2rp1q3PZnj17lJSUpNjYWGdbQECAoqKilJiYaO1Y/s6KOs5ne/fdd9WrVy/5+fm5tH/77beqUaOG6tevr/vvv9/l642QMjMz830dq2LFilq9erWys7MlSYmJiS7nqyTFxcU5z1fO6aIpylj/1cmTJ5Wdna2qVau6tHNen1tRx5rP6otzIec0n9UFS0tLk6R8P+tnc8dncWJiogIDA9WiRQtnn9jYWHl4eGjVqlVuO56/K3eMc2HbdTgcCgwMdGkfPXq0qlWrpmbNmmncuHE6ffr0hRd/CSOIQZLUqlUrzZgxQz179pS3t7dCQkIUEBCgt956y9mnfv36mjp1qj7//HN9+OGHys3N1Q033KDff/9dkpSUlCRJCg4Odtl2cHCwc9mlrijjfLbVq1dry5Yt+r//+z+X9g4dOuiDDz7QkiVLNGbMGH333Xfq2LGjcnJybBxGmRAXF6f//Oc/Wrt2rYwxWrNmjf7zn/8oOztbR48elXTmnD3X+co5XTRFGeu/evzxxxUaGuryjwLO6/MryljzWX3xintO81ldsNzcXD300ENq1aqVGjVqVGg/d3wWJyUlqUaNGi7LK1SooKpVq5b789pd4/xXp06d0uOPP67evXvL39/f2T548GDNmjVLy5Yt07333quXXnpJjz32mHsO5hJTobQLwN/Dtm3bNGTIED399NOKi4vToUOHNGzYMN1333169913JZ25mhMdHe1c54YbblBkZKTeeecdPf/886VVeplSlHE+27vvvqvGjRurZcuWLu29evVy/rlx48Zq0qSJ6tatq2+//Vbt2rUr8eMoC5566iklJSXp+uuvlzFGwcHBio+P19ixY+Xhwe+g3Km4Yz169GjNmjVL3377rctVB87r8yvKWPNZffGKe07zWV2wBx54QFu2bNEPP/xQ2qWUayUxztnZ2brjjjtkjNGkSZNclg0dOtT55yZNmsjb21v33nuvRo0aJR8fH7fVcCngXyOQdGb2nVatWmnYsGFq0qSJ4uLiNHHiRE2dOlWHDh0qcB0vLy81a9ZMv/zyiyQpJCREkpScnOzSLzk52bnsUleccc7IyNCsWbOUkJBw3u1eccUVql69uvPvAme+RjR16lSdPHlSe/fu1f79+xUeHq4qVaooKChI0plz9lznK+d00RRlrPO8/PLLGj16tBYuXKgmTZqcc7uc1/kVZ6zz8FldfMUZZz6rCzZo0CB99dVXWrZsmWrVqnXOvu74LA4JCdHhw4ddlp8+fVopKSnl+rx25zjnyQth+/bt06JFi1yuhhUkKipKp0+f1t69ey/oGC5lBDFIOnO/xl9/y+fp6SnpzJSxBcnJydHmzZtVs2ZNSVJERIRCQkK0ZMkSZ5/09HStWrXK5bezl7LijPOcOXOUmZmpu+6667zb/f3333Xs2DHn3wX+x8vLS7Vq1ZKnp6dmzZqlW265xeXKwdnnqyQtWrTIeb5yThfPucZaksaOHavnn39e8+fPd7mPozCc14U731ifjc/qC1eUceaz2pUxRoMGDdKnn36qpUuXKiIi4rzruOOzODo6WqmpqVq7dq2zz9KlS5Wbm6uoqCh3HNrfSkmMs/S/ELZr1y4tXrxY1apVO+92N2zYIA8Pj3xfDUURlMIEIbDg+PHjZv369Wb9+vVGknn11VfN+vXrnbPeDB8+3Nx9993O/tOmTTMVKlQwEydONLt37zY//PCDadGihWnZsqWzz7PPPmsWLFhgdu/ebdauXWt69eplfH19zdatW519Ro8ebQIDA83nn39uNm3aZLp161aup0QuiXHO07p1a9OzZ88C9/noo4+axMREs2fPHrN48WJz7bXXmnr16plTp06V3MGWsuKO9c6dO8306dPNzz//bFatWmV69uxpqlatavbs2ePss2LFClOhQgXz8ssvm+3bt5uRI0cWOGXypXROG1MyYz169Gjj7e1tPv74Y5cpj48fP+7cJ+e1e8aaz+r8SmKc8/BZ7er+++83AQEB5ttvv3X5WT958qSzz913322GDx/ufO+uz+IOHTqYZs2amVWrVpkffvjB1KtXr9xOX18S45yVlWW6du1qatWqZTZs2OCy3czMTGOMMStXrjSvvfaa2bBhg9m9e7f58MMPTVBQkOnbt6/dASgnCGLl1LJly4ykfK/4+HhjzJkpddu0aeOyzuuvv24aNGhgKlasaGrWrGn69Onj8oyfhx56yNSuXdt4e3ub4OBg06lTJ7Nu3TqXbeTm5pqnnnrKBAcHGx8fH9OuXTuzc+fOkj7cUlMS42yMMTt27DCSzMKFC/Pt8+TJk6Z9+/YmKCjIeHl5mTp16ph77rkn33OCypvijvW2bdtM06ZNTcWKFY2/v7/p1q2b2bFjR77tfvTRR+aqq64y3t7epmHDhubrr792WX6pndPGlMxY16lTp8Btjhw50hjDee3OseazOr+S+vzgszq/gsZZksvz1dq0aeMc+zzu+Cw+duyY6d27t6lcubLx9/c3/fv3d/6yp7wpiXHes2dPodtdtmyZMcaYtWvXmqioKBMQEGB8fX1NZGSkeemll8r1LxdKksOYQr53BgAAAAAoEdwjBgAAAACWEcQAAAAAwDKCGAAAAABYRhADAAAAAMsIYgAAAABgGUEMAAAAACwjiAEAAACAZQQxAAAAALCMIAYAKPf27t0rh8OhDRs2XNR2brrpJj300ENuqQkAcGkjiAEA3KZfv35yOBz5Xh06dCjt0qx47733Cjz+//znP27ZvrsC5cU4dOiQ7rzzTl111VXy8PAgmALABapQ2gUAAMqXDh06aNq0aS5tPj4+pVSNff7+/tq5c6dLW0BAQClVU7isrCx5e3sXe73MzEwFBQXpySef1GuvvVYClQHApYErYgAAt/Lx8VFISIjL67LLLnMuz7tCdOutt6pSpUqqV6+evvjiC5dtbN26Vbfccov8/f1VpUoVxcTEaPfu3ZKk3NxcPffcc6pVq5Z8fHzUtGlTzZ8/32X91atXq1mzZvL19VWLFi20fv36fHVu2bJFHTt2VOXKlRUcHKy7775bR48edS7PyMhQ3759VblyZdWsWVOvvPJKkY7f4XDkO/6KFSsWaZ/z589X69atFRgYqGrVqumWW25xHrckRURESJKaNWsmh8Ohm266SVLBX5ns3r27+vXr53wfHh6u559/Xn379pW/v78GDhwoSfrhhx8UExOjihUrKiwsTIMHD1ZGRkahxxceHq4JEyaob9++f8uACQBlBUEMAGDds88+qzvuuEObNm1Sp06d1KdPH6WkpEiSDhw4oBtvvFE+Pj5aunSp1q5dqwEDBuj06dOSpAkTJuiVV17Ryy+/rE2bNikuLk5du3bVrl27JEknTpzQLbfcogYNGmjt2rV65pln9Oijj7rsPzU1VW3btlWzZs20Zs0azZ8/X8nJybrjjjucfYYNG6bvvvtOn3/+uRYuXKhvv/1W69atu+BjLso+MzIyNHToUK1Zs0ZLliyRh4eHbr31VuXm5ko6EzAlafHixTp06JA++eSTYtXw8ssv65prrtH69ev11FNPaffu3erQoYN69OihTZs2afbs2frhhx80aNCgCz5OAEARGQAA3CQ+Pt54enoaPz8/l9eLL77o7CPJPPnkk873J06cMJLMN998Y4wxZsSIESYiIsJkZWUVuI/Q0FCX7RljzHXXXWf+9a9/GWOMeeedd0y1atXMn3/+6Vw+adIkI8msX7/eGGPM888/b9q3b++yjd9++81IMjt37jTHjx833t7e5qOPPnIuP3bsmKlYsaIZMmRIocc/bdo0I8nl2IODg4u0z4IcOXLESDKbN282xhizZ88el+PI06ZNm3x1devWzcTHxzvf16lTx3Tv3t2lT0JCghk4cKBL2/Lly42Hh4fL+BWmoP0CAIqGe8QAAG518803a9KkSS5tVatWdXnfpEkT55/9/Pzk7++vw4cPS5I2bNigmJgYeXl55dt2enq6Dh48qFatWrm0t2rVShs3bpQkbd++XU2aNJGvr69zeXR0tEv/jRs3atmyZapcuXK+fezevVt//vmnsrKyFBUV5XIM9evXP+exS1KVKlVcrpx5eHgUaZ9XXXWVdu3apaefflqrVq3S0aNHnVfC9u/fr0aNGp133+fTokULl/cbN27Upk2bNGPGDGebMUa5ubnas2ePIiMjL3qfAICCEcQAAG7l5+enK6+88px9/hqyHA6HM3Tk3U9Vkk6cOKEuXbpozJgx+ZbVrFlTv/zyywVv28PDo8DjP98+JalLly6qU6eOpkyZotDQUOXm5qpRo0bKyso67z6NMS5t2dnZ+fr5+fnlq+nee+/V4MGD8/WtXbv2OfcJALg4BDEAwN9KkyZN9P777ys7OztfYPP391doaKhWrFihNm3aONtXrFihli1bSpIiIyM1ffp0nTp1ynlV7Mcff3TZzrXXXqu5c+cqPDxcFSrk/19h3bp15eXlpVWrVjkDyR9//KGff/7ZZb/Fcb59Hjt2TDt37tSUKVMUExMj6cxEGmfLm+UwJyfHpT0oKEiHDh1yvs/JydGWLVt08803n7embdu2nTc4AwDcj8k6AABulZmZqaSkJJfX2TMDns+gQYOUnp6uXr16ac2aNdq1a5emT5/unBJ+2LBhGjNmjGbPnq2dO3dq+PDh2rBhg4YMGSJJuvPOO+VwOHTPPfdo27Ztmjdvnl5++WWXfTzwwANKSUlR79699dNPP2n37t1asGCB+vfvr5ycHFWuXFkJCQkaNmyYli5dqi1btqhfv37OrxleiPPt87LLLlO1atU0efJk/fLLL1q6dKmGDh3qso0aNWqoYsWKzok+0tLSJElt27bV119/ra+//lo7duzQ/fffr9TU1PPW9Pjjj2vlypUaNGiQNmzYoF27dunzzz8/72QdGzZs0IYNG3TixAkdOXJEGzZs0LZt2y54bADgUkQQAwC41fz581WzZk2XV+vWrYu8frVq1bR06VKdOHFCbdq0UfPmzTVlyhTn1bHBgwdr6NCheuSRR9S4cWPNnz9fX3zxherVqydJqly5sr788ktt3rxZzZo10xNPPJHv64B5V9VycnLUvn17NW7cWA899JACAwOdYWvcuHGKiYlRly5dFBsbq9atW6t58+YXPC7n26eHh4dmzZqltWvXqlGjRnr44Yc1btw4l21UqFBBr7/+ut555x2FhoaqW7dukqQBAwYoPj5effv2VZs2bXTFFVec92qYdObq43fffaeff/5ZMTExatasmZ5++mmFhoaec71mzZqpWbNmWrt2rWbOnKlmzZqpU6dOFzw2AHApcpi/fqkcAAAAAFCiuCIGAAAAAJYRxAAAAADAMoIYAAAAAFhGEAMAAAAAywhiAAAAAGAZQQwAAAAALCOIAQAAAIBlBDEAAAAAsIwgBgAAAACWEcQAAAAAwDKCGAAAAABY9v8Ada3SNU5nFKIAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1000x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "# Extract the encoder part of the autoencoder \n",
    "encoder_model = Model(input_layer, bottleneck) \n",
    "\n",
    "# Encode the test data \n",
    "encoded_imgs = encoder_model.predict(x_test) \n",
    "\n",
    "# Visualize the first two dimensions of the encoded features \n",
    "plt.figure(figsize=(10, 8)) \n",
    "plt.scatter(encoded_imgs[:, 0], encoded_imgs[:, 1], c='blue', alpha=0.5) \n",
    "plt.title('Encoded Features - First Two Dimensions') \n",
    "plt.xlabel('Encoded Feature 1') \n",
    "plt.ylabel('Encoded Feature 2') \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "# Extract the encoder part of the autoencoder \n",
    "encoder_model = Model(input_layer, bottleneck) \n",
    "\n",
    "# Encode the test data \n",
    "encoded_imgs = encoder_model.predict(x_test) \n",
    "\n",
    "# Visualize the first two dimensions of the encoded features \n",
    "plt.figure(figsize=(10, 8)) \n",
    "plt.scatter(encoded_imgs[:, 0], encoded_imgs[:, 1], c='blue', alpha=0.5) \n",
    "plt.title('Encoded Features - First Two Dimensions') \n",
    "plt.xlabel('Encoded Feature 1') \n",
    "plt.ylabel('Encoded Feature 2') \n",
    "plt.show() \n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Conclusion: \n",
    "\n",
    "Congratulations on completing this lab! In this lab, you have gained practical experience in building, training, and evaluating autoencoders using Keras. You have learned to preprocess data, construct a basic autoencoder architecture, train the model on the MNIST dataset, and visualize the results. Additionally, you explored fine-tuning techniques to enhance the model's performance and applied the autoencoder to denoise images. \n",
    "\n",
    "Continue experimenting with different architectures, datasets, and applications to further deepen your knowledge and skills in using autoencoders. The concepts and techniques you have learned in this lab will serve as a foundation for more advanced topics in deep learning. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Authors\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Skills Network\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright © IBM Corporation. All rights reserved.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.9"
  },
  "prev_pub_hash": "6b856929be8ddec6857f0588ee7cd695dfdfb707704a384fc8f032f2831a0ed5"
 },
 "nbformat": 4,
 "nbformat_minor": 4
}